Field and Menon 1

GLYCOLIPID ANCHORING OF CELL SURFACE PROTEINS

Mark C. Field and Anant K. Menon

Laboratory of Molecular Parasitology, The Rockefeller University, New York, NY 10021, USA.

CONTENTS I. INTRODUCTION II. DISCOVERY OF GPI 1. Release of Surface Proteins From Cells with PI-PLC 2. Analysis of Membrane Proteins Lacking Hydrophobic Peptide Anchors (Thy-1 and VSG) III. STRUCTURE OF GPI 1. Elucidation of the Structure of the VSG GPI-anchor 2. Structures of Glycosylphosphatidylinositol-Containing Species 3. Analysis of GPI Structures IV. BIOSYNTHESIS OF GPI ANCHORS 1. GPI anchor addition is a rapid post-translational event 2. Identification and Characterization of GPI Anchor Precursors 3. Assembly of GPI Anchor Precursors 4. Membrane Topology of GPI Biosynthesis 5. Sub-cellular Distribution of Non-Protein-Linked GPIs 6. Inhibitors of GPI Biosynthesis 7. Defects in GPI Anchoring 8. Protein COOH-terminal sequences signalling GPI attachment 9. Transfer of GPI precursors to protein V. BIOLOGICAL CHARACTERISTICS AND SIGNIFICANCE OF GPI ANCHORS 1. Mobility of GPI-anchored proteins: lateral diffusion in the plane of the membrane 2. Secretion, sorting, and endocytosis of GPI-anchored proteins 3. Signal transduction VI ACKNOWLEDGEMENTS VII LITERATURE CITED

Field and Menon 2

I. INTRODUCTION Over the last decade a number of new covalent modifications of proteins have been identified and characterised. Several of these involve the attachment of hydrophobic moieties to the polypeptide chain, e.g. myristoylation, isoprenylation, palmitoylation, and glypiation. The first three of these modifications are dealt with elsewhere in this volume. Glypiation,1 the attachment of a phosphatidylinositol-containing glycolipid (GPI) is perhaps the most complex in structural and biosynthetic terms. In this chapter we shall discuss several aspects of GPI biology. We shall briefly outline the data which suggested the existence of this type of protein modification, and then describe the elucidation of the structure of the GPI-anchor of the variant surface glycoprotein of (VSG) of Trypanosoma brucei. We shall also document the current structural data base of GPI (and related glycoconjugate) structures, and briefly describe the empirical approaches towards identification and elucidation of the structures of GPI species. We shall then describe the biosynthesis of GPI-anchors, and finally consider possible functional aspects of GPI-anchoring as opposed to transmembrane polypeptide-anchoring. The reader is referred to a number of recent reviews for more detailed discussions of particular aspects of this field; general overviews2,3,4, details of structure and elucidation1,5,6, summary of C-terminal extension and processing site sequences,1 possible functions7, and precursor biosynthesis8,9,10. II. DISCOVERY OF GPI 1. Release of Surface Proteins From Cells with PI-PLC Increased levels of alkaline phosphatase (AP) activity were detected in the serum of experimental animals infected with Bacillus anthrax11. The effect could be reproduced with culture filtrates from other Bacillus species, notably B. cereus and B. thuringiensis12,13 and it was proposed that the release of AP was most probably due to the action of a novel phosphatidylinositol specific phospholipase C (PI-PLC).14,15 These observations were corroborated in the 1970s by the demonstration that highly purified PI-PLCs from B. cereus, Staphylococcus aureus, and B. thuringiensis could specifically release AP from cell membranes.16,17,18 Phospholipases with other specificities were not able to duplicate this effect. The generality of the phenomenon was demonstrated when it was found that PI-PLC could release 5'-nucleotidase and acetylcholinesterase (AChE) 19,20,21 from intact erythrocytes in the absence of hemolysis. The process was specific, as certain other cell surface markers were not released.20 Characterisation of AP before and after treatment with PI-PLC demonstrated that the physical properties of the enzyme were altered. PI-PLC

Field and Menon 3

released AP was of low molecular weight, i.e. not aggregated or associated with lipid vesicles,20,21 and was unable to bind to phospholipid vesicles. Organic solvent extracted AP however could bind phospholipid,22 but this property was lost on PI-PLC treatment suggesting that the lipophilic properties of AP were intimately associated with AP itself, rather than being some property of the lipid bilayer with which it was associated, and could be explained by the covalent linkage of a PI molecule to the protein. Similar observations made for Torpedo marmorata AChE,23 coupled with analyses showing that T. marmorata AChE and AP contained myo-inositol in stoichiometric amounts,24,25 provided strong evidence of a covalent PI-protein linkage. 2. Analysis of Membrane Proteins Lacking Hydrophobic Peptide Anchors (Thy-1 and VSG) Investigations into the mechanism of anchoring of two further proteins, the rat neural and T cell antigen Thy-1 and the variant surface glycoprotein (VSG) from Trypanosoma brucei provided further evidence for a novel form of lipid-mediated anchoring. Thy-1, the smallest member of the immunoglobulin superfamily26 was considered to use an enigmatic method for membrane anchoring by virtue of the absence of a clear transmembrane region from the primary structure of the isolated glycoprotein. Thy-1 could also be released from thymocytes by PI-PLC,27 suggesting an anchor mechanism similar to that of AP and AChE. Chemical studies of the protein showed that a hydrophobic structure attached to the C-terminal cysteine contained EthN, GlcN and Gal,28 as well as myo-inositol, phosphate, glycerol, mannose and stearate.29,30 Subsequent to the identification of the complete structure of the VSG glycosylphosphatidylinositol (GPI)-anchor, Thy-1 was the first mammalian glycoprotein to have the structure of it's GPI glycan solved (see below). The structure of the lipid portion of the Thy-1 anchor remains unknown at the present time. VSG provided an excellent paradigm for the clarification of the structure of the GPI anchor, and proof that the lipid was the sole mechanism of membrane anchoring. Comparison of the complete amino acid sequence and the cDNA sequence for the VSG from T. brucei variant 11731,32 led to the prediction that hydrophobic peptide sequences from both the N and the C-terminus were removed during maturation of the glycoprotein. The N-terminal extension is the classical secretion signal sequence. The C-terminal extension (17-23 amino acids, depending on the VSG type)33 was of sufficient hydrophobicity to be a membrane insertion sequence, and was highly conserved between otherwise diverse VSG sequences.33 As isolated by cell disruption, VSG is a soluble glycoprotein (sVSG).34,35 However, lysis of trypanosomes in boiling detergent or in the presence of 10mM Zn2+ produces an amphipathic membrane form

Field and Menon 4

(mfVSG).36 Conversion of mfVSG to sVSG during osmotic shock was due to the action of an endogenous enzyme,35,37 which removed myristic acid from the VSG38 while generating the 'cross reacting determinant' (CRD), the only common immunogenic structure observed between different VSG variants.39,40 The CRD itself is associated with a complex oligosaccharide that was attached to the C-terminus of the VSG polypeptide.39,40,41,42 The linkage between the oligosaccharide and the C-terminal amino acid was shown to involve ethanolamine in amide linkage to the α-carboxyl group.43 In order for this addition to be made, removal of the C-terminal extension would have to take place. All these data came together when it was shown that the activity converting the mfVSG to sVSG was in fact an endogenous phospholipase C.44 The similarities between the biochemical data for VSG, Thy-1, AP and the other PI-PLC released enzymes suggested that a common mechanism for membrane anchoring was involved. Other proteins were also soon identified based on their susceptibility to PI-PLC, that were proposed to be GPI-anchored, e.g. Leishmania gp63.45 III. STRUCTURE OF GPI 1. Elucidation of the Structure of the VSG GPI-anchor As discussed in the previous section, data from several directions converged in the mid 1980s, and suggested that a PI containing glycolipid was responsible for the anchoring of a group of membrane glycoproteins. As an example we shall briefly discuss the elucidation of the complete structure of the VSG GPI-anchor. Ferguson and coworkers showed that the lipase activity converting mfVSG to sVSG released sn-1,2-dimyristoylglycerol from the mfVSG.44 The C-terminal domains from all VSGs, generated by protease treatment, contained mannose, ethanolamine, glucosamine and varying amounts of galactose.46 Unusually, the glucosamine was not N-acetylated.47 Nitrous acid treatment of the sVSG did not release any carbohydrate (which cleaves glycan chains by deamination of non-N-acetylated amino-sugars), suggesting that the glucosamine was located at the reducing terminus of the glycan.48 It was then shown that deamination of 3H-myristate-labelled mfVSG released PI,48 and that treatment with S. aureus PI-PLC released 100% of the radiolabel, demonstrating that all fatty acid substituents were contained within the PI moiety. Composition analysis of the C-terminal glycopeptide following treatment with mild acid and alkaline phosphatase identified myo-inositol in this VSG fragment, as well as galactose and mannose.48 The inability to detect inositol in previous work was ascribed to the unusual acid stability of the GlcN-Inositol linkage. Recovery of inositol-1-phosphate following alkaline phosphatase treatment of the deaminated glycopeptide, suggested formation of a cyclic inositol phosphate product from the PI-PLC cleavage,

Field and Menon 5

which was confirmed by mild acid treatment.48 This feature is important in generation of the CRD epitope.49 The complete glycan structure for VSG117 and the linkage of ethanolamine-phosphate to the glycan was determined by chemical analysis, utilising nuclear magnetic resonance (NMR), gas chromatography mass spectrometry (GCMS), chemical and enzymatic degradation (Figure 1). The C-terminal glycopeptide was prepared from mfVSG by pronase digestion, and a neutral glycan prepared by deamination, reduction and dephosphorylation.50 Proton NMR and methylation analysis provided linkage positions, anomericity and composition, whilst gel filtration and exoglycosidase treatments provided an estimate of glycan microheterogeneity as well as confirmation of the NMR/GCMS assignments. Periodate oxidation and GCMS of the reaction products was used to provide data on the Man-GlcN-Inositol linkage positions. The residue to which EtN-P was linked was determined by the resistance of the mannose core of the glycan to α-mannosidase digestion, and subsequent sensitivity following dephosphorylation. Similar approaches have been used to deduce the structures of additional GPI-anchors, e.g. acetylcholinesterase,51,52 gp63 (PSP) from Leishmania major53 and the glycan from the Saccharomyces cerevisiae 125kDa glycoprotein (C. Fankhauser, personal communication), although in the former case more extensive characterisation by fast atom bombardment mass spectrometry (FABMS) was utilised. The structure of the GPI-anchor of the scrapie form of the prion protein PrPSc has been determined by a novel approach utilising electrospray mass spectrometry and laser desorption mass spectrometry54 but these data have not been confirmed by an independent method. Proteins with a GPI anchor are listed in Table 1. 2. Structures of Glycosylphosphatidylinositol-Containing Species Many workers have described the structures of a variety of GPI-containing molecules. These essentially fall into three classes, the GPI-protein anchors (Tables 1 and 2), the lipophosphoglycans and glycosylinositolphospholipids (LPGs, GIPLs, Table 3) and the free GPI lipids (Table 2), some of which are protein anchor precursors. Protein GPI-Anchors; All protein-linked GPI anchors defined to date contain a consensus core glycan, which is conserved to the level of linkage position. Table 2 summarises known GPI-anchor structures and the structures of homologous free GPI-lipids, which are presumed to be biosynthetic intermediates in the construction of the protein anchor. Apart from conservation of the glycan core sequence, and, by definition, the presence of PI, the GPI-anchors represent a diverse set of stuctures.

Field and Menon 6

Lipid Structures; Diacyl,44,48,50 monoacyl,55 alkylacyl,51,53 and ceramide-based PIs (Fankhauser, personal communication)56 have been reported as PI structures of protein GPI-anchors. So far, lyso-alkyl PI has only been reported for the lipophosphoglycan of leishmanial parasites (see below).57 In addition, the presence of a fatty acid substituent on the inositol ring has been reported for a number of proteins and GPI-anchor precursors, including protozoan antigens (PARP/procyclin)55 and GPI precursors58,59,60 and mammalian proteins (acetylcholinesterase)51 and precursors.61,62 The inositol fatty acid is palmitate in all cases so far determined, which is in marked contrast to the heterogeneity displayed by the rest of the PI moiety (see Table 2). The functional significance of palmitoyl-inositol is not understood, but an important consideration is that the presence of this modification renders the GPI-anchor resistant to cleavage by bacterial PI-PLCs,51,55,59,60 and therefore will not allow release of proteins linked to the cell membrane by such anchors, and also will not allow the generation of the CRD epitope. As both of these observations are important criteria for the detection of a GPI, it is important to be aware that such variants of GPI-anchor structure exist, and may be quite common (see section III). Interestingly, the GPI-anchor precursors recently identified in T cells are PI-PLC resistant,61,63 probably due to inositol acylation, whilst the anchors of the surface glycoproteins are PI-PLC sensitive, e.g. Thy-1 (see biosynthesis, section IV.3). Glycan Decorations; A number of additional substituents have been described that are added to the GPI core. It is believed that many of these substituents are added after initial transfer of the glycolipid to the protein (see section IV.3, Table 2). As with structural microheterogeneity in other glycosylation reactions, i.e. N and O-linked glycans, modification of the core is non-stoichiometric. For example, substitution of the VSG 117 GPI-anchor with α-galactose produces four major structures, containing 2 to 4 α-Gal residues,50 with the Gal3 species consisting of two isomers. Other minor species are probably also present. Different VSG classes contain either no galactose or up to 8 galactose residues.5,46 For Thy-1, the fourth mannose residue is present in only 80% of the structures recovered from the protein.63 The functional significance of any of the additional substituents has not been determined. Factors controlling maturation of the GPI-anchors have also not been assessed. From the types of substituents present it is most likely that the majority of these additions take place within the Golgi compartment. In the case of VSG, GPI-anchor processing occurs with a half-life of approximately 15 minutes,65 at a similar time as the processing of the N-linked glycans to an endoglycosidase H resistant form, implying a cis or medial Golgi location. However, our knowledge of the compartmentalisation of the trypanosome secretory pathway is rudimentary and therefore extrapolation to the mammalian system is not completely valid.

Field and Menon 7

Lipophosphoglycan and Glycoinositolphospholipids; In Leishmania spp. (a group of South American trypanosomatids) a second class of GPIs have also been identified66,67,68,69 and extensively characterised. Structural information and citations for this class of molecules are summarised in Table 3. These lipids (GIPLs) are homologous to the protein-linked GPIs and their precursors, but identity is restricted to the core sequence Manα1-4GlcNα1-6PI. Some GIPLs are expressed at the cell surface and are immunogenic.67 GIPLs may serve as precursors for a major cell surface glycoconjugate termed lipophosphoglycan (LPG)57,70,71,72,73 that may be involved in the infection of macrophages by Leishmania.74,75 These organisms also synthesise the protein anchoring GPI series.53,76 Interestingly, the presence of an acyl-inositol-PI has not been seen in the Leishmania GIPL/LPG series, and therefore this feature of GPI decoration may be restricted to the protein-anchor series, or these organisms may lack the enzyme neccessary for addition of the acyl moiety to inositol. Species specific expression of different GIPLs has been reported. In L. major, the GIPLs contain terminal αGal residues, whereas in L. donovani the GIPLs are structurally more closely related to the GPI-protein anchor precursors identified in other organisms.69 In addition the structures of the glycan headgroups of the L. donovani lipids are developmentally regulated, with a Manα1-3 only being seen in the promastigote stage and not in the amastigote (Table 3). Interestingly, expression of LPG is also shut off in the amastigote stage.69 A large glycoconjugate, lipopeptidophosphoglycan, (LPPG), with a similar structure to LPG, has been reported in Trypanosoma cruzi (Table 3).77,78 The LPPG structure is more closely related to the protein-anchor series, but interestingly contains galactose-furanose substituents, which have also been seen in LPG. Little is known about the biosynthesis and function of this species beyond its structure. The precise relationship of the LPG/LPPG/GIPL series to the protein-linked GPI series remains to be fully evaluated and they will not be discussed further. 3. Analysis of GPI Structures Methodology for the study of GPI-anchors and their precursors has been well documented over the past few years, and most information can be obtained in the original papers themselves. Given below is a brief overview of the methods that are available, and the information that can be obtained by each one. The reader is directed to other articles,79,80 and a forthcoming volume in the 'Practical Approach' series published by IRL Press on 'Lipid Modifications of Proteins' (Hooper and Turner, eds) for detailed protocols. Essentially the methods can be considered as three units, (i) identification of a GPI-anchored protein or lipid, (ii) dealing with the protein

Field and Menon 8

and releasing the GPI moiety or extracting and purifying the lipid, and (iii) analysing the glycan headgroup, and lipid moiety. Identification of GPI-Anchored Proteins: Methods depend on either composition, incorporation of GPI components in metabolic-labelling experiments, or on the sensitivity of the protein to enzyme cleavage. The methods used will depend on factors such as abundance, availability of a rigorous purification protocol, and of monospecific antibodies. Composition analysis for ethanolamine can be performed using an amino acid analyser, and is a useful method available to most laboratories. For inositol or fatty acid analysis, GCMS is required, which necessitates greater investment both financially and in terms of time. Composition data are only valid for extensively purified material, which in most cases will not be available. Metabolic-labelling with ethanolamine or inositol is informative as these compounds are not incorporated into many other protein-linked structures, and therefore is quite specific. Fatty acid labelling (usually palmitate or myristate) coupled with another criterion is also useful, but many proteins are acylated which do not contain GPI-anchors (see other chapters in this volume). Incorporation of sugars is less specific as the labelling must be performed in the presence of tunicamycin (to inhibit addition of N-linked sugars), but inhibition may be incomplete or the protein could be O-glycosylated, complicating analysis. Treatment with N-glycanase or endoglycosidase H can be used to discriminate between GPI and N-linked glycosylation. Under some circumstances labelling with particular compounds will be unsuccessful for obscure reasons, for example in T. brucei, GPI lipids and proteins are not labelled with [3H]-inositol, whereas S. cerevisiae and Leishmania GPIs are labelled efficiently.66,81,82 Use of inhibitors, e.g., D-mannosamine or PMSF, may also be helpful (section IV.6). These compounds prevent synthesis of GPI-anchor precursors, and can be used to inhibit the incorporation of 3H-ethanolamine into GPI-anchored proteins (see section IV.6).83,84 Alteration of biochemical properties by treatment with bacterial PI-PLC is a powerful method for the detection of GPI-anchored proteins. Incubation of intact cells with PI-PLC can result in the selective release of GPI-anchored proteins into the culture medium. Fluorescent activated cell sorter analysis of endogenous GPI-anchored proteins requires the availability of good antibodies, but is an extremely powerful method. This procedure is also of use in demonstrating GPI-anchoring of cloned protein sequences that have potential GPI-anchor addition signals that can be expressed in heterologous eukaryotic systems. Commonly, products of PI-PLC digests are analysed by detergent-phase partitioning, using TX-114.85,86 PI-PLC sensitive GPI-anchored proteins are converted from amphipathic proteins (which are recovered in the detergent-enriched phase) to hydrophilic proteins (recovered in the aqueous phase). Samples can then be fractionated by gel

Field and Menon 9

electrophoresis. In addition, PI-PLC generates the CRD,49 which contains a cyclic 1,2-inositol-phosphate, generated by the action of the enzyme. Therefore, immunoprecipitation or western-blotting following PI-PLC treatment can be used as a further demonstration of the presence of a GPI-anchor. The CRD epitope is destroyed by mild acid treatment, which can be incorporated as an additional control for the assay.53 It should be noted that the structure of the CRD epitope is complex, so that negative reactions (when there is other information supporting the possibility of GPI-anchorage) should be treated with caution.49 A very real problem with PI-PLC cleavage is the occurence of acyl-inositol. This modification will prevent PI-PLC action. One possible method is to perform a partial base treatment, to remove some of the acyl groups from the inositol. Alternatively the protein could be treated with mammalian serum GPI-PLD. In order for this enzyme to work the protein of interest must be fully solubilised, as incorporation of the protein into a lipid bilayer prevents the action of the enzyme.87a In this case, the PLD releases a phosphatidic acid from the GPI-anchor, which can be analysed by thin-layer chromatography.51,55 Obtaining a GPI Fragment from an Intact Glycoprotein: The most obvious and simplest method to produce a GPI fragment is exhaustive proteolysis, with, for example, pronase.50,53 This method works in most cases to produce a C-terminal glycopeptide, containing a one or a few amino acids. The hydrophobic GPI can be purified by extraction into organic solvent and further processed. In the case of a glycoprotein containing a convenient protease site, a specific protease can be used to release the GPI (e.g. papain cleavage of AChE).87 The amino acid composition of the C-terminal glycopeptide, or even N-terminal sequencing can be used at this stage to determine the GPI attachment site. The glycan can at this point be released by deamination and reduction, to cleave off the PI moiety, and HF treatment to remove the ethanolamine phosphate. A radiolabel can be introduced by reduction with 3H-sodium borohydride. Direct analysis of the GPI moiety by mass spectrometric techniques can be utilised, as exemplified for acetylcholinesterase.52 Alternatively the protein can be treated with PI-PLC at the begining of the procedure. Differential sensitivity of the glycan to exoglycosidase digestion before and after release from the residual peptide and ethanolamine-phosphate can be used to determine the position of attachement of the phosphate to the core glycan.50 Purification of Lipid Species: For GPI-anchor precursors, which are free lipids, removal of the protein is eliminated. However, the glycolipids must be extracted and purified from the cells of interest. This is most conveniently performed by using metabolically labelled material, as the quantity of GPI lipids is small compared to cellular phospholipid. Standard extraction protocols with chloroform methanol-containing solvents are suitable.79,80,88 Subsequent fractionation and purification is most easily achieved by thin-layer chromatography, but other methods involving Iatrobead silica liquid-

Field and Menon 10

chromatography80,89 or octyl-sepharose66 can also be used. The glycan can then be released following the same procedures as described for proteins. Analysing the GPI-Glycan: The methodology employed will depend on the amount of material available. For plentiful samples (>50nmoles), NMR and mass spectrometry analysis are the methods of choice. As several GPI-glycans have already been studied, interpretation of NMR spectra is now somewhat easier. Composition and linkage analysis by GCMS, followed by exoglycosidase and chemical degradation will provide additional information, allowing a complete assignment to be made in most cases. At present, the number of structures that have been sequenced is too small to allow complete reliance on exoglycosidase digestion and chromatographic properties for a full assignment, but this may become possible in the near future. Analysis of the precursors (with the exception of the GIPLs from Leishmania) relies on metabolically-labelled material. There are now a large number of GPI lipids which have been characterised. Essentially, headgroup analysis is performed by liquid chromatography (most commonly utilising the Dionex anion exchange HPLC system) coupled with exoglycosidase digestion of the deaminated glycan.90 Gel filtration and ion exchange are also useful additional methods to consider. Linkage analysis has been performed on some of the T. brucei precursors by methylation of the glycan before hydrolysis.90,91,92 Because of the smaller amounts of material available, analysis of free GPIs will remain less rigorous than that of the protein-linked GPI-anchors themselves. Analysis of Lipid Substituents: For the GPI-anchored proteins, preliminary information can be obtained by treatment of the protein with PI-PLC, GPI-PLD, or sphingomyelinase.56 Analysis of both the released fragment and the efficiency of release are important in assigning the architecture of the GPI-anchor. For example, an inositol-acylated anchor releases only a fraction of lipid following GPI-PLD digestion.55,60 Analysis of fragments can be performed by GCMS or by TLC with radiolabelled material. Base hydrolysis can be used to determine the presence of ester-linked fatty acids, and phospholipase A2 to probe the individual fatty acid substituents. Fatty acid composition data can be obtained by conversion of the fatty acid to a methylester, followed by fractionation by TLC or GCMS. Alkyl-acyl structures are usually analysed as fatty acid methylesters and methylalkylglycerols.53 IV. BIOSYNTHESIS OF GPI ANCHORS 1. GPI anchor addition is a rapid post-translational event Studies from many laboratories have shown that GPI anchor addition is signalled by a COOH-terminal amino acid sequence that is rapidly removed upon completion of protein synthesis and replaced by a preformed GPI

Field and Menon 11

(Figure 3). The rapidity of COOH-terminal processing and GPI addition was first discovered in pulse-chase experiments with T. brucei which showed that VSG was already GPI-anchored after a short pulse (1 min) and chase (0 and 2 min).65,93,94 In JEG-3 human choriocarcinoma cells GPI addition to AP occurs within 5 minutes of protein synthesis.95 Rapid addition of GPI to protein has also been demonstrated for Thy-1 glycoprotein,96 carcinoembryonic antigen (CEA),97 and the neural cell adhesion molecule, N-CAM-120.98 In addition to the rapid kinetics of processing, other data indicate that GPI addition occurs early in the secretory pathway, probably in the endoplasmic reticulum. (i) GPI anchoring proceeds normally in the sec18 secretion mutant of Saccharomyces cerevisiae when vesicular transport between the ER and Golgi is blocked at the non-permissive temperature.99 (ii) Newly synthesized decay accelerating factor (DAF) in HeLa cells can be metabolically labelled with [3H]ethanolamine before O-linked glycans are added100 indicating that GPI addition occurs before the protein enters the Golgi apparatus. (iii) GPI-anchored proteins can be isolated with immature N-linked glycans, indicating that the GPI anchor is added prior to N-linked glycan processing in the Golgi.65,94,95,97 (iv) COOH-terminal cleavage and GPI-addition activity can be detected in ER microsomal preparations from tissue culture cells101,102,103 and from canine pancreas (J. Vidugiriene, unpublished observations). 2. Identification and Characterization of GPI Anchor Precursors African Trypanosomes; The rapid processing of proteins to GPI-anchored forms suggested that the GPI anchor was pre-assembled. This proposal was stengthened when a polar lipid, P2 (also called Lipid A,104 or Glycolipid A89), isolated from extracts of bloodstream form T. brucei metabolically labelled with [3H]ethanolamine, [3H]mannose, [3H]glucosamine, [3H]myristic acid, or [32P], was shown to have the essential structural features of a GPI anchor.104,105 Metabolic labelling of P2 was not affected by tunicamycin, a potent and specific inhibitor of the first step in the synthesis of dolichol-linked oligosaccharides.90,106 Labelling in the presence of cycloheximide, a protein synthesis inhibitor, showed that P2 was still synthesised, indicating that the lipid was not produced by adventitious cleavage of GPI-anchored VSG.105 Estimates based on inositol analysis90 and metabolic incorporation of radioactive precursors of defined specific activity105 indicated a pool size of ~104 molecules of P2 per cell, compared with ~107 molecules of VSG per cell. P2 was characterized as ethanolamine-P-Manα1-2Manα1-6Manα1-GlcN-Inositol-P-sn 1,2-dimyristylglycerol by a variety of methods59,90,105 (see section III). The linkages not defined in the structure are likely to be

Field and Menon 12

identical to those in the VSG GPI-anchor since glycan fragments prepared from metabolically labelled P2 were found to be indistinguishable from similar preparations derived from the core of the VSG GPI anchor (Table I). However, unlike the VSG-linked GPI, the ethanolamine residue in P2 contains a free amino group.105 A second polar lipid termed P3105 (also called Lipid C or Glycolipid C)89,104 was also found in bloodstream form trypanosome extracts. P3 was less polar than P2, but, like P2, could be metabolically labelled via [3H]ethanolamine, [3H]mannose, [3H]glucosamine, [3H]myristic acid, and [32P]. Although P3 appeared to be a GPI by compositional criteria, it was not cleaved by PI-PLC, a primary diagnostic of GPI structures. Subsequent to the elucidation of a PI-PLC-resistant inositol acylated GPI-anchor structure for human acetylcholinesterase,51 P3 was shown to be identical to P2 except for the presence of palmitic acid esterified to one of the inositol hydroxyl groups.58,59 P3 could be converted to P2 by selective removal of the palmitate residue.59 The significance of P3 in trypanosomes is the subject of some speculation and will be discussed later in this chapter (section IV.9). The GPI anchors of trypanosome VSGs are galactosylated to different extents, the average number of galactose residues per anchor depending on the particular trypanosome variant (table 2). Addition of galactose residues to the VSG anchor appears to occur in the Golgi apparatus.65 Galactosylated forms of P2 and P3, structurally homologous to the galactosylated VSG GPI anchor, are also synthesized by bloodstream trypanosomes.107 Curiously, P2 galactosylation is extensive but only the monogalactosylated form of P3 has been detected. More recently, GPI lipids have also been identified in insect-stage (procyclic) T. brucei, and in a variety of mammalian cells and yeast. The single abundant GPI species (PP1)60 identified in procyclic trypanosomes is a variant of P3: it is PI-PLC resistant and contains one palmitic acid residue esterified to inositol, and one stearic acid residue esterified to the 1-position of glycerol (Table 2). An identical arrangement of fatty acids was subsequently found in the GPI anchor of the major procyclic surface protein55 consistent with PP1 as the GPI anchor precursor in procyclic trypanosomes. Furthermore, PP1 is synthesised from a diacyl structure, homologous to P3.108 Other systems; GPI anchor precursors in mammalian cells are being studied in several laboratories, and, although the structural data are incomplete, some general points may be made. Contrary to original expectations based on the abundance of GPI-anchored proteins in mammalian cells versus T. brucei, the mammalian GPIs appear to be about as abundant as the trypanosome GPI precursors. In a variety of mammalian cells, the most abundant [3H]ethanolamine-labelled GPI was PI-PLC resistant and more polar than trypanosome P2. The GPIs were inositol acylated and most probably contained one or two extra phosphoethanolamine moieties

Field and Menon 13

attached to the mannose residue proximal to glucosamine.61,62 Side-chain phosphoethanolamine moieties have been described in the GPI anchors of Thy-1 and AChE64,52 and their presence on GPI anchor precursors indicates that they are added to the GPI structure early, rather than as a late event during transit of the protein through the secretory pathway. An alternative explanation is that a large proportion of the metabolically labelled GPIs are exported to the cell surface as well, and are modified with phosphoethanolamine side-chains en route (section IV.5). Curiously, only PI-PLC-resistant GPI anchor precursors were found in the cells examined, even though the cells expressed PI-PLC-sensitive GPI-anchored proteins. This observation is consistent with the hypothesis that many GPI-anchored proteins are initially PI-PLC-resistant and aquire sensitivity following removal of the inositol-linked fatty acid later in the secretory pathway before the protein reaches the cell surface. 3. Assembly of GPI Anchor Precursors Glycan Biosynthesis; Much of our present information on GPI anchor biosynthesis is derived from investigations in the trypanosome system, beginning with the demonstration of the cell-free synthesis of P2 and P3.89,91,109 Metabolic labelling experiments indicated that tunicamycin had no effect on GPI biosynthesis,59,90,94 which allowed the use of this inhibitor prior to lysis89 and in the labelling reactions in order to suppress labelling of lipid intermediates in the N-glycosylation pathway. It was clear that the radiolabelled sugar nucleotides UDP-[3H]GlcNAc and GDP-[3H]mannose could be used as sugar donors in vitro to generate [3H]P2, [3H]P3, and a range of [3H]glycolipids characterized as GlcNAc-PI, GlcN-PI, Man1-3GlcN-PI, and Man2-3GlcN-acylPI.89,91,110 Although the identification of partially constructed GPI species suggested that, like dolichol-PP-oligosaccharide biosynthesis,111,112 GPI assembly involved sequential addition of monosaccharides to lipid acceptors (Figure 4), it remains possible that the assembly pathway is more complex. The first committed step in GPI assembly is the transfer of GlcNAc from UDP-GlcNAc to PI to form GlcNAc-PI. GlcNAc-PI is then rapidly de-acetylated to form GlcN-PI.91,110,113,114 Three murine T cell mutants (complementation classes A, C, and H) originally selected for their inability to express cell surface Thy-1115,116 are in fact defective in GlcNAc-PI synthesis.63,113 Since each of the three mutants belongs to a separate complementation class, it is clear that three genes are directly involved in GlcNAc-PI synthesis. One explanation is that the glycosyltransferase responsible for GlcNAc-PI synthesis is a three-subunit enzyme, and all three subunits are required for function. An alternative possibility is that another protein involved in GlcNAc-PI synthesis, e.g. the UDP-GlcNAc transporter, responsible for

Field and Menon 14

transport of UDP-GlcNAc from the cytosol into the lumen of the ER111,112,112a is defective. It remains to be established whether GlcNAc-PI is synthesized in the lumenal leaflet of the ER membrane, and therefore, if a defect in UDP-GlcNAc transport would affect GPI assembly (see below). Although GDP-mannose is able to support GPI assembly in vitro,89,91,109 it is not the direct donor of the three core mannose residues in the GPI structure. Dolichol-P-mannose (DPM), a lipid metabolite synthesized in the ER from GDP-mannose and dolichol-P, was implicated as the mannose donor for at least one if not all the GPI mannose residues when a Thy-1 negative murine T cell mutant (Class E)115,116 known to be defective in DPM synthesis117,118 was shown to be unable to synthesize the Thy-1 GPI anchor.119,120,121,122 Also GPI-anchoring is defective in yeast mutants generated by disruption of the dolichol-P-mannose synthase DPM1 gene,82 and defective Thy-1 and Ly-6A surface expression in a DPM synthase deficient T cell hybridoma mutant (M38/20)123 could be restored by transfecting the cells with the yeast DPM1 gene.61 The involvement of DPM in GPI synthesis was also suggested by experiments showing that a DPM synthesis inhibitor 2-deoxy-2-fluoroglucose (FGlc, see section IV.6), could suppress the synthesis of P2 and P3 in bloodstream trypanosomes.124 FGlc has been used more recently to inhibit the synthesis of GPI-anchored alkaline phosphatase in tissue culture cells.125 Similar data were obtained from in vitro labelling experiments using GDP-[3H]mannose: incorporation of [3H]mannose into P2 and P3 was stimulated by dolichol-phosphate and inhibited by amphomycin, an antibiotic inhibitor of DPM synthesis92 (see section IV.6). The role of DPM in GPI assembly was established firmly by direct in vitro labelling using dolichol-P-[3H]mannose, followed by methylation linkage analysis of neutral glycans derived from labelled P2, Man2GlcN-PI, and Man3GlcN-PI. The results demonstrated conclusively that the assembly of all three mannose residues in the GPI core required DPM.92 These data also raised the possibility that elongation may not occur by sequential transfer of mannose residues9,92 and were consistent with a scheme in which Man1GlcN-PI is elongated by the direct transfer of Manα1-2Man from a DPM-derived donor, bypassing Man2GlcN-PI. The Man2GlcN-PI species detected in the labelling may be adventitiously derived by α-mannosidase trimming of Man3GlcN-PI. However, these biosynthetic alternatives remain to be established. Phosphoethanolamine; The terminal phosphoethanolamine residue in the GPI structure appears to be derived from phosphatidylethanolamine (PE).10,126,127 Direct experimental proof of phosphoethanolamine transfer from PE to a Man3 GPI acceptor has been difficult to obtain, but several lines of evidence are consistent with this model. In vitro labelling with CDP-

Field and Menon 15

[3H]ethanolamine and [β-32P]CDP-ethanolamine shows that phosphoethanolamine is transferred as a group, and in vitro pulse-chase studies clearly rule out CDP-[3H]ethanolamine as the direct donor of the GPI phosphoethanolamine moiety. Furthermore, metabolic labelling of trypanosomes with [3H]serine labels GPIs from radiolabelled PE synthesized via decarboxylation of labelled phosphatidylserine (PS). Related experiments with yeast mutants unable to synthesize PE from CDP-ethanolamine due to disruption of the genes encoding the ethanolaminephosphotransferases show that, when incubated with [3H]ethanolamine, the mutant cells can synthesize [3H]ethanolamine-phosphate and CDP-[3H]ethanolamine but not [3H]PE, and they also cannot incorporate [3H]ethanolamine into GPI anchors. However the ability of the mutant cells to synthesize GPI anchors, as assessed by [3H]inositol labelling, is unaltered because they make PE by decarboxylation of PS. Phosphatidylinositol Specificity; In bloodstream trypanosomes, the PI molecular species selected for GPI biosynthesis contains stearic acid in the 1-position and an as yet undefined fatty acid in the 2-position of the glycerol moiety.128 Once the GPI structure is assembled the glycerol-linked fatty acids are sequentially replaced with myristic acid.89,128 Re-modelling of the Man3 GPI intermediate can occur if phosphoethanolamine addition is inhibited.84 PP1, the GPI anchor precursor in procyclic stage trypanosomes contains only one glycerol-linked stearic acid at the analogous position to the unremodelled bloodstream form GPI, and is the consequence of partial re-modelling of a diglyceride-containing precursor.108 Similar fatty acid re-modelling reactions might occur in other systems giving rise to unique molecular species (glycerol fatty acids) in GPI anchors that are not representative of the molecular species spectrum found in total cellular PI. However, it is difficult to see how this process could occur with ether-linked fatty acids found in many GPI anchors (table 1). Only a minor subset of the spectrum of PI molecular species available in cells of the electric organ of the electric eel Torpedo marmorata, appears in the GPI anchor of Torpedo acetylcholinesterase.129 These data could be explained by postulating fatty acid re-modelling reactions, or by suggesting that only a particular subset of PI molecular species is available to the first GPI glycosyltransferase. Analysis of the PI labelled with [3H]fatty acid in procyclic trypanosomes suggests that the GPI species are derived from a minor fraction of the cellular PI pool.55 The inability to label GPIs via [3H]inositol in trypanosome lysates91 or in culture (Field and Menon, unpublished) despite efficient labelling of PI, suggests that the PI required for GPI assembly is a very minor fraction of the labelled PI. This may reflect targetting of specific molecular species of PI to the domains of the ER, e.g., the inner leaflet of the ER membrane bilayer, where GPI biosynthesis is initiated.

Field and Menon 16

Inositol Acylation; The biosynthetic pathway outlined above ignores the relationship between PI-PLC-resistant (i.e., inositol acylated) GPI species, and PI-PLC-sensitive forms. The role of inositol acylation in GPI assembly is not understood, and it remains to be established whether inositol acylation merely represents a GPI structural variation, or whether it is biosynthetically important. Since the original characterization of P3 in bloodstream trypanosomes, several inositol acylated GPI intermediates have been detected in extracts from in vitro labelling reactions using trypanosome membrane preparations. Although analyses of GPI biosynthetic intermediates in bloodstream trypanosomes show that a only minor proportion are inositol acylated, similar experiments with procyclic stage trypanosomes which express a major inositol acylated protein55,130 show that the degree of inositol acylation increases with the size of the GPI glycan chain to the point where the mature phosphoethanolamine-containing GPI anchor precursor, PP1, is exclusively inositol acylated.60 The smallest inositol acylated GPI intermediate detected was Man1GlcN-acylPI in procyclics.108 In other experiments in bloodstream trypanosomes, GlcN-acylPI may also have been synthesized91 suggesting that the inositol-linked fatty acid is added very early in the assembly pathway. Pulse-chase experiments in procyclic trypanosomes suggest that the inositol-linked fatty acid is not derived directly from fatty acyl Coenzyme A, but instead from a relatively metabolically stable donor such as a phospholipid.55 However these results need to be re-examined in the context of recent data indicating that inositol acylation in yeast membrane preparations is dependent on fatty acyl CoA.131 In all cases where the fatty acid esterified to the inositol has been identified, it has been shown to be palmitic acid (Table 2). In the case of T. brucei inositol acylation in the insect-dwelling form may be important to guard against the effects of PI-PLC on surface antigens as certain Bacillus species are insect pathogens.132 Studies of GPI biosynthesis in mammalian cells provide a clearer indication of the potential functional significance of inositol acylation. The analyses have been facilitated by the availability of a series of T-cell mutants which are blocked at different stages of GPI biosynthesis, and which therefore accumulate GPI biosynthetic intermediates.115,116,123 Possibly unlike the sequence of events in bloodsteam trypanosomes, inositol acylation in mammalian cells occurs at the GlcN-PI stage and persists through all the subsequent mannosylated structures, including the GPI anchor precursor.61,62,63,133,134 Thus the initial assembly sequence is GlcNAc-PI-> GlcN-PI-> GlcN acylPI. GlcN-acylPI accumulates in the absence of GPI mannosylation, and is therefore easily detectable by metabolic labelling in mammalian and yeast mutants that are defective in DPM synthesis.82,131,134 With the exception of GlcNAc-PI and GlcN-PI, no PI-PLC-sensitive (non-inositol-acylated) GPI intermediates or anchor precursors have been detected in wild-type T-cells or in a spectrum of T-cell mutants suggesting that

Field and Menon 17

inositol acylation is important for mannosylation. This effect may be a consequence of the profound stereochemical constraint imposed on the GPI structure by the inositol-linked fatty acid.10,91 A number of reports are consistent with a hypothesis that GPI anchor precursors are initially PI-PLC-resistant and are processed to a PI-PLC-sensitive form after transfer to protein. As discussed above, GPI anchor precursors in a variety of mammalian cells appear to be exclusively PI-PLC-resistant, even though GPI-anchored proteins in the same cells are PI-PLC-sensitive suggesting either conversion of the GPI from a PI-PLC-resistant to a sensitive form after attachment to protein, or rapid processing of the precursor just prior to transfer. In two sublines of human K562 cells, mature AChE is either PI-PLC-sensitive or resistant, but in both lines AChE that has not been sialylated is PI-PLC-resistant, suggesting that all newly synthesized AChE contains an inositol acylated GPI anchor, which is subsequently processed during transit to the cell surface in one of the two cell lines.135 PI-PLC sensitivity or resistance of GPI-anchored proteins appears to be cell-type and not protein specific. In several murine L cell lines, PI-PLC-resistance of GPI-anchored proteins behaves as a recessive trait in cell fusion experiments,136 consistent with the presence of an inositol-specific acyl hydrolase in the cells with PI-PLC-sensitive GPI-anchored proteins. The sub-cellular location of the inositol-specific acyl hydrolase remains to be determined, as does its specificity for GPI-anchored-proteins versus non-protein-linked GPIs. However, the 110kDa precursor of the yeast gp125 is PI-PLC sensitive, even when studied in sec 18 cells at the non-permissive temperature (a block in ER to cis-Golgi trafficking), suggesting that, if GPI-anchors are added as inositol-acylated forms, then deacylation must occur in the ER compartment.99 This topic is discussed further elsewhere.137 4. Membrane Topology of GPI Biosynthesis The model that GPI anchors are constructed via the sequential glycosylation of PI is further complicated by membrane topological issues. The end products of the assembly process are presumed to be associated with the luminal leaflet of the ER in order to be transferred to a newly synthesized polypeptide. However, the starting material, PI, is synthesized from CDP-diacylglycerol and inositol on the cytoplasmic face of the ER138 suggesting that at some stage of the assembly process a cytosolically disposed lipid intermediate flips to the luminal face of the ER. There are several other indications that GPI assembly is likely to involve flipping of intermediates between the cytoplasmic and luminal leaflets of the ER accompanied by the transport of cytosolically synthesized precursors into the lumen of the ER. The three components required for the synthesis of GPI anchors from PI are glucosamine, mannose, and phosphoethanolamine. The most immediate (identified) donors of these forms of all three are available

Field and Menon 18

on both faces of the ER membrane. UDP-GlcNAc can be transported into the ER,111,112,112a whilst synthesis of DPM from cytosolic GDP-Man followed by translocation of the lipid into the luminal leaflet111,112,139 provides a source of activated mannose on both sides of the ER. Also the phosphoethanolamine donor, PE, is available in both leaflets of the ER,140 so that no indication of the step at which the presumed translocation of the nascent GPI occurs can be gleaned from donor location. 5. Sub-cellular Distribution of Non-Protein-Linked GPIs There is no direct information concerning the sub-cellular location of GPI biosynthesis. As discussed above, the lipids are presumed to be assembled in the ER and located in the lumenal leaflet of the ER membrane bilayer in order to be transferred to newly translated and translocated proteins. However, some observations suggest that a proportion of the anchor precursors may be found in other sub-cellular locations, indicating that they are exported from the ER after synthesis. (i) Pulse-chase labelling of procyclic trypanosomes with [3H]glucosamine showed that ca. 60% of labelled PP1 was resistant to the chase, and could still be detected after several hours. The most plausible explanation is that this fraction corresponds to PP1 that is no longer in the ER, and therefore unavailable for transfer to protein. However, as it is still recoverable as PP1 no processing of the majority of the lipid occured.55 (ii) Galactosylated forms of P2 and P3 are readily labelled in vitro, via UDP-[3H]galactose, in the absence of other sugar nucleotides, rather than via GDP-[3H]mannose (in the presence of non-radioactive UDP-galactose), implying galactosylation of pre-existing rather than in vitro de novo synthesized P2 and P3, probably in a different sub-cellular compartment.107 Since galactosylation of the VSG GPI anchor apparently occurs as the protein traverses the Golgi apparatus en route to the cell surface,65 P2 and P3 are most likely galactosylated in the Golgi. More elaborate GPI lipid glycans have also been detected in the procyclic form trypanosome.60 (iii) GPI related molecules (primarily GIPL-2) have been detected on the surface of Leishmania major promastigotes using specific antibodies66,67 and are immunogenic to their mammalian hosts. 6. Inhibitors of GPI Biosynthesis

Several compounds have been shown to affect GPI biosynthesis in vivo and in vitro (Figure 4). Except PMSF all affect other types of glycosylation as well. The inhibitors and their properties are summarized below. 2-fluoro-2-deoxy-D-glucose (FGlc) - FGlc prevents the synthesis of GPI anchors by inhibiting the formation of DPM (but not GDP-mannose) in vivo.124,125,141,142,143,144 The mechanism of the DPM synthesis inhibition is unknown, but it is clear that FGlc (like other fluorosugar inhibitors of

Field and Menon 19

glycosylation) has to be metabolized in order to exert its effect, and therefore will not function in cell free assays. It should be noted that FGlc treatment has to be carried out in cell culture medium that is essentially glucose-free, and that the inhibitor (used at ~1mM) affects protein synthesis (considerable inhibition in bloodstream trypanosomes, ~30% inhibition in JEG-3 choriocarcinoma cells125). Amphomycin - Amphomycin, a lipopeptide antibiotic, forms a complex with dolichol-P and prevents it from participating in cell-free reactions leading to the synthesis of DPM, dolichol-P-glucose, and dolichol-PP-GlcNAc.145,146 As with FGlc, GPI synthesis is blocked because DPM synthesis is inhibited.92 Amphomycin inhibition can be overcome by adding excess dolichol-P.92 Mannosamine (2-amino-2-deoxy-D-mannose) (ManN) - ManN blocks the formation of α1,2 mannose linkages in N-linked sugars,147 and inhibits GPI synthesis in vivo possibly by inhibiting the α1,2 mannosyl transferase responsible for the conversion of Man2- to Man3-GPI intermediates, or by acting as a chain terminator.83 Although the mechanistic details remain to be elucidated, it is known that ManN is ineffective in cell-free systems since it probably needs to be converted to an active metabolite in order to act (Field and Menon, unpublished data).147 Also, unlike FGlc, ManN has very little effect on protein synthesis at concentrations at which it effectively inhibits GPI biosynthesis. ManN also is able to inhibit biosynthesis of the LPG species in Leishmania (Medina-Acosta and Field, unpublished data). Phenylmethanesulphonyl fluoride (PMSF) - The serine esterase inhibitor, PMSF, suppresses incorporation of the terminal phospho-ethanolamine moiety into the GPI chain, resulting in the accumulation of a Man3 GPI species.84 Inhibition can be achieved both in vitro and in vivo. The same effect can also be achieved in vitro with diisopropyl fluorophosphate, a reagent which phosphorylates activated serines.84 7. Defects in GPI Anchoring Several GPI biosynthesis mutants have been described in the previous sections. Many of these are rare naturally occurring mutants selected from populations of wild-type cells. Others were selected after mutagenesis of wild-type cells.123,148,149,150 Commonly used L-cell fibroblast cell lines are unable to synthesize GPI-anchored proteins for reasons that remain to be defined.151,152 Chinese hamster ovary (CHO) cell line (B421) with a minimal capacity to synthesize DPM are also defective in GPI anchoring; however, the defect can be corrected by exposing the cells to tunicamycin153 which presumably diverts the residual DPM from the N-glycosylation pathway to the GPI pathway.

Field and Menon 20

Mutant thymoma cell lines derived from rare Thy-1- cells isolated originally by using cytotoxic immunoselection procedures on Thy-1+ T cell lymphomas adapted to tissue culture115,116 have been mapped into nine complementation classes. Six of the mutant classes define genes acting at points in the GPI assembly pathway62,63,113,116,117,118,119,121,122,154: classes A, C, and H do not synthesize GlcNAc-PI (see above), class E cells cannot synthesize DPM and therefore accumulate GlcN-acylPI, class B GPI synthesis is blocked at the Man2 stage, class F cells cannot add the terminal phosphoethanolamine group and are therefore blocked at the Man3 stage (class F cells also have an ether lipid defect which appears to be unrelated to GPI synthesis)155. For unknown reasons, the aminoglycoside G418 can reversibly induce class F and B cells to express some cell surface Thy-1.156 The human disease, paroxysmal nocturnal hemoglobinuria (PNH) correlates with defects in GPI-anchor synthesis. PNH is an acquired hemolytic anemia in which variable proportions of circulating blood cells, arising from clonal expansion of a somatically mutated pluripotent stem cell, show increased sensitivity to autologous complement lysis owing to the absence of the GPI-anchored complement regulatory proteins DAF157,158 and homologous restriction factor.159 In addition, the cells lack a number of other GPI anchored proteins including LFA-3,160 AP,161 CD14,162 and CD16.163 The absence of several GPI-anchored proteins coupled with the failure of affected cells to express cell-surface DAF despite the presence of a normal DAF gene, mRNA transcripts, and nascent DAF polypeptides164,165 indicates that PNH is due to a defect(s) in some stage of GPI assembly or transfer. The precise identification of the defect has been hindered by the absence PNH-derived cell lines; however cloned natural killer cells carrying the PNH defect have recently been described.165a 8. Protein COOH-terminal sequences signalling GPI attachment Signals for GPI Addition; A recognition/signalling event is clearly involved in directing attachment of GPI anchors to some proteins and not others, since only a restricted set of cell surface proteins are GPI-anchored. Indications of the nature of the GPI attachment signal came originally from comparisons of the cDNA-predicted and actual amino acid sequences of a trypanosome VSG and Thy-1 (see section II). Both the proteins had been processed at the N- and at the C-terminus.26,28,31,32,166 The N-terminal processing was identified as the normal removal of an N-terminal signal sequence by ER signal peptidase, after translocation of the polypeptide into the lumen of the ER. The cleaved, relatively hydrophobic C-terminal sequence is now known to be the signal for GPI attachment. Transfection experiments using a variety of gene constructs, showed that addition or substitution of the C-terminal sequence could convert water-soluble or

Field and Menon 21

transmembrane proteins to GPI-anchored forms in mammalian cells.151,152,67,167,167a,168,169,170,171 Dissection of the C-terminal signal sequence by construction of chimeras and site-specific mutants, showed that the signal was composed of two distinct elements, a short (generally 15-20 residues) hydrophobic domain, and a processing site consisting of 2-3 small amino acid residues positioned 10-12 amino acids N-terminal to the hydrophobic domain (Figure 3). Although the hydrophobic domain had to be of a certain length, the specific sequence was unimportant as it could be replaced with a random hydrophobic peptide without affecting GPI anchoring.169,172,173 Curiously, the hydrophobic signal was found to be effective even when positioned at an internal location rather than at the C-terminus.174 Site-specific mutagenesis of the cleavage/attachment site of DAF and placental AP, and identification of the attachment site in a number of other proteins indicated that only small amino acids (Ser, Gly, Ala, Asp, Asn, Cys) would function at the attachment site.1,175,176 Small amino acids are also generally found in the +1 and +2 positions relative to the attachment site, i.e., the first two amino acids of the cleaved C-terminal peptide. Mutagenesis experiments by Gerber et al.177 showed that while a spectrum of substitutions at the +1 site were permissible (Asp, Ser, Cys, Met, Thr, Glu; reduced processing with Trp; no processing with Pro) the requirements at the +2 site were far more stringent. Of the 11 amino acid substitutions examined at the +2 site, only Gly and Ala permitted normal processing (although only trace activity was detected with Ser or Cys in these assays, it should be noted that Ser occurs naturally in the +2 position in trypanosome VSGs sequenced to date !). In separate experiments with a series of chimeras consisting of sequences from DAF, the LDL receptor, and human growth hormone, the positioning of the small amino acid domain relative to the hydrophobic stretch appeared to be critical since GPI anchoring efficiency dropped sharply if the spacing was changed from the optimum of 10-12 amino acids.178 However, these experiments may need to be examined further since the precise attachment site was not determined for the processed mutants. Differential Splicing/Expression; Proteins such as LFA-3179 and N-CAM180,181,182,183 are naturally expressed in both GPI-anchored and transmembrane forms on the same cell. The different forms probably arise from a single gene by differential processing of the primary mRNA transcript. Other mechanisms for the expression of transmembrane versus GPI-anchored forms also exist. Alternate forms of CD16 [Fcγ receptor III] have been found on different cell types through cell-specific expression of similar but distinct genes, rather than alternative mRNA processing.184 9. Transfer of GPI precursors to protein

Field and Menon 22

Despite rapid GPI-anchoring of newly synthesized VSG in trypanosomes, membranes prepared from bloodstream trypanosomes contain a small amount of non-glycolipid-anchored VSG which can accept glycolipid precursors synthesized in situ or added exogenously.185 Although the membranes synthesize roughly equal quantities of radiolabelled P2 and P3, VSG labelled under the conditions of the in vitro assay is exclusively PI-PLC-sensitive, indicating that only P2 is transferred to the protein. The membranes also catalyze the transfer of exogenously added GPIs to VSG, demonstrating that the trypanosome GPIs P2, P3, and PP1, are all bona fide GPI anchor precursors. Transfer does not require ongoing protein synthesis and efficiency is unaltered in the presence of non-hydrolyzable ATP analogues or an ATP regenerating system,185 consistent with a transpeptidation reaction mechanism (see below). Curiously, when radiolabelled P3 is added to the membranes, the processed (labelled) VSG is found to be PI-PLC-resistant185 suggesting a lack of specificity under the assay conditions that is not reflected in living cells or when the precursors are synthesized in situ in membranes. In these situations P2 may have a competitive advantage over P3 resulting in exclusively PI-PLC-sensitive VSG. Alternatively, both lipids may be transferred to VSG acceptors but P3-anchored VSG may be rapidly inositol-deacylated to give PI-PLC-sensitive VSG, and this process is not supported by the in vitro system. Non-GPI-anchored protein precursors can be detected in mammalian cells, particularly in mutants incapable of synthesizing complete GPI structures119,120,154 and in normal cells after inhibition of GPI biosynthesis with FGlc125 or ManN.83 The unprocessed proteins could not be detected in cells metabolically labelled with [3H]ethanolamine or [3H]fatty acids, consistent with the lack of a GPI anchor. The proteins clearly retained the C-terminal hydrophobic tail since they partitioned into the detergent phase in Triton X-114 phase partitioning experiments, and unprocessed Thy-1 in five thymoma mutants (Classes A, C, E, F and H) could be metabolically labelled with [3H]tryptophan (the only tryptophan residue in Thy-1 is located in the C-terminal tail).119,154 In the absence of GPI synthesis, the nascent proteins may persist for several hours before finally being cleaved at or near the GPI attachment site, and secreted into the medium.83,125 Cells of the Class B mutant secrete Thy-1 after removal of some or all of the C-terminal domain.154 However, some thymoma mutants (Classes A, B, C, and E), apparently accumulate uncleaved Thy-1 (detected in metabolic labelling over a few hours) with immature N-linked glycans, suggesting that the peptide tail impairs exit from the ER or cis-Golgi.119,154 Curiously, in cells where GPI synthesis is not completely suppressed, but nevertheless reduced considerably to the point where detection of GPI anchor precursors is difficult, efficiency of transfer of GPI anchors appears to be selective so that some proteins are processed to a greater degree than others. For example in a T-cell hybridoma clone M43/8 which synthesizes

Field and Menon 23

anchor precursors at about 1% the level of wild-type, the surface expression of GPI-anchored Ly-6A was ~2% of wild-type, whereas Thy-1 was ~20%.186 Similar results were obtained from analyses of neutrophils from patients with PNH: expression of GPI-anchored CD16 (Fcγ receptor III) expression was preserved, although reduced, whereas other GPI-anchored proteins (DAF (CD55) and CD59) were completely absent.187 It has been difficult to demonstrate the transfer of GPIs to proteins using mammalian in vitro translation-translocation systems, although the C-terminal cleavage event has been clearly described and indirect proof of glycolipid-anchoring has been obtained.101,102,103,188,189 The C-terminally processed protein was at least partly susceptible to PI-PLC.103 One difficulty has been to obtain GPI precursors of sufficiently high specific activity for use in in vitro assays where only femtomole amounts of protein are translated and processed. Nevertheless, considerable progress has been made, and it is clear that the GPI-anchoring event can be reproduced and studied in vitro. Recent experiments have used an engineered protein (derived from placental AP) that is of sufficiently low molecular weight so as to easily detect molecular weight changes caused by N- and C-terminal processing events.102 When the protein was synthesized in vitro in the presence of CHO cell microsomes, C-terminal processing lagged behind N-terminal signal sequence cleavage. C-terminal cleavage did not occur if microsomes derived from a GPI-mutant cell line (M38/20, defective in dolichol-P-mannose synthesis) were used instead, but was restored when polar lipids (presumably including GPI anchor precursors) from wild-type cells were added to the mutant cell microsomes (Kodukula, Yeh and Udenfriend, unpublished data). The in vitro system described provides a basis for elucidating details of the GPI anchoring event, and possibly identification of the enzyme(s) involved. Mechanistically, GPI addition to protein appears to be a transpeptidation reaction possibly analogous to the penicillin-sensitive crosslinking of peptidoglycan strands during the synthesis of bacterial cell walls.190 The amino terminus of a pentaglycine side-chain on one peptidoglycan strand attacks the terminal peptide bond in another strand to crosslink the two strands and release the terminal D-Alanine.191,192 The reaction sequence includes formation of an R-D-Ala-enzyme complex, followed by transfer of R-D-Ala to the amino terminus of glycine. The reaction does not require an external source of energy since the energy of the D-Ala-D-Ala bond is conserved in the R-D-Ala-enzyme complex. Transpeptidation reactions are also carried out by a variety of peptidases193,194 and γ-glutamyl transpeptidases.195 Recently described examples of protein splicing196,197,198 may also involve transpeptidation, but in these cases definite mechanistic proof remains to be established. V. BIOLOGICAL CHARACTERISTICS AND SIGNIFICANCE OF GPI ANCHORS

Field and Menon 24

1. Mobility of GPI-anchored proteins: lateral diffusion in the plane of the membrane Lateral diffusion coefficients of membrane-spanning proteins are typically substantially lower than those for lipids (membrane proteins D ~10-10 cm2/s, lipids D ~10-8 cm2/s) presumably because of constraints imposed on protein diffusion by interactions with the extracellular matrix and the cytoskeleton.199,200 These interactions presumably also account for the immobile fraction (usually 20-50%, D < 10-12 cm2/s) of protein molecules consistently detected in fluorescence photobleaching recovery (FPR) measurements. Since the GPI anchor does not cross the membrane bilayer, GPI-anchored proteins might be expected to be relatively highly mobile compared to membrane-spanning proteins, and possess a small, possibly zero, immobile fraction. However, diffusion measurements for several GPI anchored proteins show that the GPI-anchor does not necessarily confer a high degree of mobility, and that mobility may be determined primarily through interactions of the extracellular domain with other cell surface components.201 It should be noted that the GPI-anchored folate receptor is highly clustered and presumably immobile in the plasma membrane of epithelial cells, a state dependent on cholesterol and therefore attributed to lipid-lipid rather than protein-protein interactions.202,203 One difference between otherwise similar GPI-anchored versus transmembrane proteins (such as the Class I major histocompatibility molecules H-2Db (membrane-spanning) and Qa-2 (GPI-anchored)) is that membrane-spanning proteins appear to be confined to micrometer scale membrane domains defined by dynamic cytoplasmic barriers which do not restrict the long range (several micrometers) lateral movement of GPI-anchored proteins.204,205 2. Secretion, sorting, and endocytosis of GPI-anchored proteins The plasma membrane of polarized epithelial cells is divided into two membrane domains (apical and basolateral) by a circumferential ring of tight junctions formed between adjacent cells. The unique composition of each plasma membrane domain is apparently achieved, at least in some cell types (such as the Madin-Darby canine kidney (MDCK) cell line), by packaging of apical and basolateral components into separate carrier vesicles derived from the trans-Golgi network (TGN).206,207 GPI anchored proteins (such as AP, folate receptor, trehalase, and aminopeptidase P) are located preferentially on the apical surface of polarized epithelia,202,208,209,210 suggesting that the GPI anchor may contain apical targeting information.208,212,213,214 This possibility was explored in transfection experiments in MDCK cells using a variety of fusion proteins. Conversion of normally basolaterally targeted membrane-spanning or secreted proteins to

Field and Menon 25

GPI-anchored forms resulted in apical expression of the proteins in MDCK cells.208,215 These results were extended in other experiments in which the GPI signal in AP was replaced with the transmembrane-cytoplasmic domain of a normally basolaterally targeted membrane protein resulting in basolateral expression.208 Inhibition of GPI synthesis with ManN (see above) resulted in accumulation of an unprocessed protein precursor which was eventually cleaved and secreted without polarity.83 Although the GPI anchor may compete with other - possibly conflicting or redundant - targeting signals in the ectodomains of GPI-anchored proteins it appears to act as a “dominant” apical sorting signal in renal (e.g., MDCK) and intestinal (e.g., CaCo-2) epithelial cell lines.212,213 Interestingly, polarized sorting of GPI-anchored proteins also occurs in neurons: Thy-1 is expressed on axonal membranes but not on the neuronal cell bodies or dendrites.216 The mechanism by which GPI-anchored proteins are delivered to the apical surface is not known. The most attractive hypothesis207 is that GPI-anchored proteins cluster with glycosphingolipids and relevant accessory proteins in the cytosolic leaflet of the TGN membrane. These patches eventually form vesicles containing the necessary markers and machinery for specific delivery to the apical plasma membrane. Evidence for such aggregates has come from analyses of changes in detergent solubility of GPI-anchored proteins (specifically AP) as they traverse the secretory pathway.217 AP became insoluble in ice-cold Triton X-100 before reaching the TGN, whereas transmembrane proteins were completely solubilized. The detergent insoluble material was enriched in sphingolipids relative to glycerophospholipids, consistent with the notion that GPI-anchored proteins and glycolipids aggregate in the Golgi as a first step toward sorting and delivery to the apical plasma membrane. GPI-anchored proteins appear to be excluded from the clathrin-mediated endocytic pathway. Thy-1, Ly-6A, 5’-nucleotidase, and the folate receptor are poorly represented in clathrin coated pits when compared with transmembrane proteins such as the transferrin receptor and the LDL receptor.202,218,219,220 The folate receptor mediates the transport of 5-methyltetrahydrofolic acid into the cytoplasm by using an unusual non-clathrin-coated pit pathway.202,203,221 The receptor is distributed in remarkably dense clusters in the apical plasma membrane of a kidney epithelial cell line (MA104), typically near non-clathrin coated membrane invaginations called caveolae.221 Clustering diminishes in the absence of cholesterol.203 Rothberg et al.202,222 propose that the receptor mediates 5-methyltetrahydrofolic acid transport into the cytoplasm by partitioning into the caveolae which are capable of transiently closing leaving the receptor inaccessible to antibody probes. The closed caveolar compartment may be acidifed to release the folate, which then enters the cytoplasm via a transmembrane transporter present in the caveolae. The folate is trapped in

Field and Menon 26

the cytoplasm after being covalently modified. The caveolae then unseal and the receptors become available for another cycle. 3. Signal transduction A number of GPI-anchored proteins are implicated in signalling events in T lymphocytes and in other hematopoietic cells.223 Activation of T cells resulting in lymphokine production and cell proliferation can be achieved by antibody-mediated cross-linking of GPI-anchored proteins such as Thy-1,224 Qa-2,225 or Ly-6A (also known as TAP).123,226 Activation via GPI-anchored proteins seems to require a functional T-cell receptor, since mutant cell lines deficient in the receptor cannot respond to antibodies against Thy-1 or Ly-6.227 Transmembrane forms of Qa-2 and Ly-6A/E cannot mediate T-cell activation,225,226 suggesting that the GPI anchor is a critical component of the stimulus. The precise role of GPI-anchored proteins in T-cell signalling is far from clear, and it is not possible at this stage to reconcile all the available data. Recent results indicate that GPI-anchored proteins are not required in the early stages of T-cell activation leading to production of interleukin-2 (IL-2), since a T-cell hybridoma mutant defective in GlcNAc-PI synthesis responded normally to mitogenic stimuli.228 An explanation for these apparently conflicting results may lie in the differences between normal T-cells and T-cell hybridomas, particularly in the requirement for cellular adhesion for activation. Removal of the majority of GPI-anchored proteins from the T-cell surface by PI-PLC treatment eliminates the ability of the cells to respond to mitogens such as Concanavalin A,229,230 despite apparently undiminished binding of fluorescent Con A to the cell surface. Curiously, activation of T-cells rendered the GPI-anchored proteins resistant to PI-PLC,229,231,232 possibly because of protective interactions with other proteins or membrane components necessary for or produced by the activation process. PI-PLC treatment may remove a protein critical for T-cell-accessory cell interaction, or, possibly eliminate critical high affinity Con A binding sites. In this context it should be noted that a T-cell hybridoma mutant expressing only intermediate affinity Con A binding sites (high affinity sites corresponding to high mannose glycans were not synthesized owing to a dolichol-P-mannose synthase defect) was shown to be unresponsive to antigen, superantigen, and Con A.228 What, if any, is the role of the GPI-anchor in cell signalling? It is possible that the signalling mechanism involves the phospholipase-mediated degradation of the GPI anchor resulting in diacylglycerol which could flip into the inner leaflet of the plasma membrane bilayer and activate protein kinase C. In this context it is interesting to note the identification of a PI-specific phospholipase C activity at the surface of Swiss 3T3 cells.233 Proteolytic or other cleavage of the anchor, coupled with the action of a

Field and Menon 27

phospholipase, could result in a phosphoglycan moiety with second messenger properties as proposed for the insulin system2,234 - the glycan could be delivered to the cytosol via a caveola-based mechanism as proposed for the delivery of folate.222 Interestingly, early investigations of the effect of B. anthrax suggested that the toxin (subequently identified as PI-PLC) had insulinomimetic properties.12 Alternatively, GPI-anchored proteins may mediate activation signals through an interaction between the GPI structure and one or more membrane proteins, such as the T-cell receptor-CD3 complex.223 Interactions between GPI anchors and membrane proteins have been suggested previously. For example, 10-20% of cell-surface Thy-1 was shown to be in a detergent-insoluble complex involving a protease-sensitive 55kDa membrane protein, minor amounts of Qa-2 and a variety of membrane associated proteins that could not be labelled by surface iodination.235 Recent data describing co-immunoprecipitation of the cytosolic protein tyrosine kinase p56lck with some GPI-anchored proteins in human and mouse T-cells,236 provides another example of a GPI anchor-protein interaction. Again, interaction of the GPI anchor with p56lck must be mediated by one or more intermediary proteins since the kinase is a myristylated cytosolic protein and the GPI is confined to the opposite (external) leaflet of the plasma membrane bilayer. VI ACKNOWLEDGEMENTS This work was supported by NIH Grants AI21531 (to G.A.M. Cross) and AI28858 (to A.K.M.). We thank Drs. K. Kodukula, M.G. Low, M. Soloski, V. Stevens, S. Udenfriend, j. Vidugiriene, S. Vijaysarathy, and E.T.H. Yeh for helpful discussions and comments on the manuscript, and Dr M.G. Low for invaluable assistance in compiling Table 1.

Field and Menon 28

VII LITERATURE CITED 1. Cross, G.A.M. Glycolipid anchoring of plasma membrane proteins.

Ann. Rev. Cell Biol. 6, 1, 1990. 2. Low, M.G., and Saltiel, A. Structural and functional roles of glycosyl-

phosphatidylinositol in membranes. Science 239, 268, 1988. 3. Low, M.G. The glycosyl-phosphatidylinositol anchor of membrane

proteins. Biochim. et Biophys. Acta 988, 427, 1989. 4. Low, M.G. Glycosyl-phosphatidylinositol: a versatile anchor for cell

surface proteins. FASEB J. 3, 1601, 1989. 5. Ferguson, M.A.J., and Williams, A.F. Cell surface anchoring of proteins

via glycosylphosphatidyl-inositol structures. Ann. Rev. Biochem. 57, 285, 1988.

6. Thomas, J.R., Dwek, R.A., and Rademacher, T.W. Structure,

biosynthesis and function of glycosylphosphatidyl-inositols. Biochemistry 29, 5413,1990

7. Lisanti, M.P., Rodriguez-Boulan, E., and Saltiel, A.R. Emerging

functional roles for the glycosyl-phosphatidylinositol membrane protein anchor. J. Memb. Biol. 117, 1, 1990.

8. Doering, T.L., Masterson, W., Hart, G.W., and Englund, P.T.

Biosynthesis of glycosylphosphatidyl-inositol membrane anchors. J. Biol. Chem. 265, 611, 1990.

9. Field, M.C., and Menon, A.K. Biosynthesis of glycosylphosphatidyl-

inositol membrane anchors in Trypanosoma brucei. Trends Glycosci. Glycotech. 10, 111, 1991.

10. Menon, A.K. Biosynthesis of glycosyl-phosphatidylinositol. Cell Biol.

Int. Rep. 15, 1007, 1991. 11. Smith, R., Keppie, J., Stanley, J.L., and Harris-Smith, P.W. The chemical

basis of the virulence of Bacillus anthrax. IV: Secondary shock as the major factor in the death of guinea-pigs from anthrax. Br. J. Exp. Pathol. 36, 323, 1955.

12. Slein, M.W., and Logan, G.F. Mechanism of action of the toxin of

Bacillus anthracic. I. Effect in vivo on some blood serum components. J. Bacteriol. 80, 77, 1960.

Field and Menon 29

13. Slein, M.W., and Logan, G.F. Mechanism of action of the toxin of

Bacillus anthracic. II. Alkaline phosphatasemia produced by culture filtrates of various Bacilli. J. Bacteriol. 83, 359, 1962.

14. Slein, M.W., and Logan, G.F. Partial purification and properties of two

phospholipases of Bacillus cereus. J. Bacteriol. 85, 369, 1963. 15. Slein, M.W., and Logan, G.F. Characterization of the phospholipases of

Bacillus cereus and their effects on erythrocytes, bone, and kidney cells. J. Bacteriol. 90, 69, 1965.

16. Low, M.G., and Finean. J. Release of alkaline phosphatase from

membranes by a phosphatidylinositol-specific phospholipase C. Biochem. J. 167, 281, 1977.

17. Ikezawa, H., Yamanegi, M., Taguchi, R., Miyashita, T., and Ohyabu, T.

Studies of phosphatidylinositol phosphodiesterase (phospholipase C type) of Bacillus cereus. I. Purification, properties and phosphatase-releasing activity. Biochim. Biophys. Acta 450, 154, 1976.

18. Taguchi, R., Yoshinari, A., and Ikezawa, H. Purification and properties

of phosphatidylinositol-specific phospholipase C of Bacillus thuringiensis. Biochim. Biophys. Acta 619, 48, 1980.

19. Taguchi, R., and Ikezawa, H. Phosphatidylinositol-specific

phospholipase C from Clostridium novyi type A. Arch. Biochem. Biophys. 186, 196, 1978.

20. Low, M.G., and Finean, J. Non-lytic release of acetylcholinesterase

from erythrocytes by a phosphatidylinositol-specific phospholipase C. FEBS Lett. 82, 143, 1977.

21. Low, M.G., and Finean, J. Specific release of plasma-membrane

proteins by a phosphatidylinositol-specific phospholipase C. Biochim. Biophys. Acta 508, 565, 1978.

22. Low, M.G., and Zilversmit, D.B. Role of phosphatidylinositol in

attachment of alkaline phosphatase to membranes. Biochemistry 19, 3913, 1980.

23. Futerman, A.H., Fiorini, R.M., Roth, E., Low, M.G., and Silman, I.

Physicochemical behaviour and structural characteristics of membrane-bound acetylcholinesterase from Torpedo electric organ.

Field and Menon 30

Effect of phosphatidylinositol-specific phospholipase C. Biochem. J. 226, 369, 1985.

24. Futerman, A.H., Low, M.G., Ackermann, K.E., Sherman, W.R., and

Silman, I. Identification of covalently bound inositol in the hydrophobic membrane-anchoring domain of Torpedo acetylcholinesterase. Biochem. Biophys. Res. Comm. 129, 312, 1985.

25. Low, M.G., Futerman, A.H., Ackermann, K.E., Sherman, W.R., and

Silman, I. Removal of covalently bound inositol from Torpedo acetylcholinesterase and mammalian alkaline phosphatases by deamination with nitrous acid. Biochem. J. 241, 615, 1987.

26. Williams, A.F., and Gagnon, J. Neuronal cell Thy-1 glycoprotein:

homology with immunoglobulin. Science 216, 696, 1982. 27. Low, M.G., and Kincade, P.W. Phosphatidylinositol is the membrane

anchoring domain of the Thy-1 glycoprotein. Nature (Lond.) 318, 62, 1985.

28. Campbell, D.G., Gagnon, J., Reid, K.B.M., and Williams, A.F. Rat brain

Thy-1 glycoprotein: the amino acid sequence, disulphide bonds and an unusual hydrophobic region. Biochem. J. 195, 15, 1981.

29. Tse, A.G.D., Barclay, A.N., Watts, ,A., and Williams, A.F. A

glycophospholipid tail at the carboxyl terminus of the Thy-1 glycoprotein of neurons and thymocytes. Science 230, 1003, 1985.

30. Williams, A.F., and Tse, A.G.D. A glycophospholipid covalently

attached to the C-terminus of the Thy-1 glycoprotein. Biosci. Reports 5, 999, 1985

31. Allen, G., Gurnett, L.P., and Cross, G.A.M. The complete amino acid

sequence of a variant surface glycoprotein (VSG 117) from Trypanosoma brucei. J. Mol. Biol. 157, 527, 1982.

32. Boothroyd, J.C., Paynter, C.A., Coleman, S.L., and Cross, G.A.M.

Complete nucleotide sequence of complementary DNA coding for a variant surface glycoprotein from Trypanosoma brucei. J. Mol. Biol. 157, 547, 1982.

33. Cross, G.A.M. Structure of the variant glycoproteins and surface coat

of Trypanosoma brucei. Phil. Trans. Roy. Soc. Lond. Ser. B. 307, 3, 1984.

Field and Menon 31

34. Cross, G.A.M. Identification, purification and properties of variant-specific glycoprotein antigens constituting the surface coat of Trypanosoma brucei. Parasitology 71, 393, 1975.

35. Cross, G.A.M. Release and purification of Trypanosoma brucei variant

surface glycoprotein. J. Cell Biochem. 24, 79, 1984a 36. Cardosa de Almeida, M.L., and Turner, M.T. The membrane form of

variant surface glycoproteins of Trypanosoma brucei. Nature 302, 349, 1983

37. Cardosa de Almeida, M.L., Allan, L.M., and Turner, M.T. Purification

and properties of the membrane form of variant surface glycoproteins (VSGs) from Trypanosoma brucei. J. Protozool. 31, 53, 1984.

38. Ferguson, M.A.J., and Cross, G.A.M. Myristylation of the membrane

form of a Trypanosoma brucei variant surface glycoprotein. J. Biol. Chem. 259, 3011, 1984.

39. Holder, A.A., and Cross, G.A.M. Glycopeptides from variant surface

glycoproteins of Trypanosoma brucei: C-terminal location of antigenically crossreacting carbohydrate moieties. Mol. Biochem. Parasitol. 2, 135, 1981

40. Holder, A.A. Characterisation of the cross-reacting carbohydrate

groups on two variant surface glycoproteins of Trypanosoma brucei. Mol. Biochem. Parasitol. 7, 331, 1983.

41. Barbet, A.F., and McGuire, T.C. Cross-reacting determinants in

variant-specific surface antigens of African trypanosomes. Proc. Natl. Acad. Sci. USA 75, 1989, 1978.

42. Cross, G.A.M. Crossreacting determinants in the C-terminal region of

trypanosome variant surface antigens. Nature 277, 310, 1979. 43. Holder, A.A. Carbohydrate is linked through ethanolamine to the C-

terminal amino acid of Trypanosoma brucei variant surface glycoprotein. Biochem. J. 209, 261, 1983.

44. Ferguson, M.A.J., Low, M.G., and Cross, G.A.M. Glycosyl-sn-1,2-

dimyritylphospatidylinositol is covalently linked to Trypanosoma brucei variant surface glycoprotein. J. Biol. Chem. 260, 14547, 1985.

Field and Menon 32

45. Bordier, C., Etges, R.J., Ward, J., Turner, M.J., and Cardoso de Almeida, M.L. Leishmania and Trypanosoma surface glycoproteins have a common glycophospholipid membrane anchor. Proc. Natl. Acad. Sci. USA 83, 5988, 1986.

46. Holder, A.A. Glycosylation of the variant surface antigens of

Trypanosoma brucei. Curr. Top. Microbiol. Immunol. 117, 57, 1985 47. Strang, A.M., Williams, J.M., Ferguson, M.A.J., Holder, A.A., and Allen,

A.K. Trypanosoma brucei brucei variant surface glycoprotein contains a non-N-acetylated glucosamine. Biochem. J. 234, 481, 1986.

48. Ferguson, M.A.J., Haldar, K., and Cross, G.A.M. Trypanosoma brucei

variant surface glycoprotein has a sn-1, 2-dimyristyl glycerol membrane anchor at its COOH terminus. J.Biol. Chem. 260, 4963, 1985.

49. Zamze, S.E., Ferguson, M.A.J., Collins, R., Dwek, R.A., and Rademacher,

T.W. Characterisation of he cross-reacting determinant (CRD) of he glycosyl-phosphatidylinositol membrane anchor of Trypanosoma brucei variant surface glycoprotein. Eur. J. Biochem. 176, 527, 1988.

50. Ferguson, M.A.J., Homans, S.W., Dwek, R.A., and Rademacher, T.W.

Glycosyl-phosphatidylinositol moiety that anchors Trypanosoma brucei variant surface glycoprotein to the membrane. Science 239, 753, 1988.

51. Roberts, W.L., Myher, J.J., Kuksis, A., Low, M.G., and Rosenberry, T.L.

Lipid analysis of the glycoinositol phospholipid membrane anchor of human eryhtrocyte acetylcholinesterase: palmitoylation of inositol results in resistance to phosphatidylinositol-specific phospholipase C. J. Biol. Chem. 263, 18766, 1988.

52. Roberts, W.L., Santikarn, S., Reinhold, V.N., and Rosenberry, T.L.

Structural characterization of the glycoinositol phospholipid membrane anchor of human erythrocyte acetylcholinesterase by fast atom bombardment mass spectrometry. J. Biol. Chem. 263, 18776, 1988.

53. Schneider, P., Ferguson, M.A.J., McConville, M.J., Mehlert, A., Homans,

S.W., and Bordier, C. Structure of the glycosylphosphatidylinositol membrane anchor of Leishmania major promastigote surface protease. J. Biol. Chem. 265, 16955, 1990

Field and Menon 33

54. Baldwin, M.A., Stahl, N., Reinders, L.G., Gibson, B.W., Prusiner, S.B., and Burlingame, A.L. Permethylation and tandem mass sepctrometry of oligosaccharides having free hexosamine: analysis of the glycoinositol phospholipid anchor glycan from the scrapie prion protein. Anal. Biochem. 191, 174, 1990.

55. Field, M.C., Menon, A.K., and Cross, G.A.M. A

glycosylphosphatidylinositol protein anchor from procyclic stageTrypanosoma brucei: lipid structure and biosynthesis. The EMBO J. 10, 2731, 1991.

56. Stadler, J., Keenan, T.W., Bauer, G., and Gerisch, G. The contact site A

glycoprotein of Dictyostelium discoideum carries a phospholipid anchor of a novel type. The EMBO J. 8, 371, 1989.

57. Orlandi, PA., and Turco, S.J. Structure of the lipid moiety of the

Leishmania donovani Lipophosphoglycan. J. Biol. Chem. 262, 10384, 1987

58. Krakow, J.L., Doering, T.L., Masterson, W.J., Hart, G.W., and Englund,

P.T. A glycolipid from Trypanosoma brucei related to the variant surface glycoprotein membrane anchor. Mol. Biochem. Parasitol. 36, 263, 1989.

59. Mayor, S., Menon, A.K., and Cross, G.A.M. Glycolipid precursors for

the membrane anchor of Trypanosoma brucei variant surface glycoproteins: I. Lipid structures of the phosphatidylinositol-specific phospholipase C sensitive and resistant glycolipids. J. Biol. Chem. 265, 6174, 1990.

60. Field, M.C., Menon, A.K., and Cross, G.A.M. Developmental variation

of glycosylphosphatidylinositol membrane anchors in Trypanosoma brucei: identification of a candidate biosynthetic precursor of the glycosylphosphatidylinositol anchor of the major procyclic stage surface glycoprotein. J. Biol. Chem. 266, 8392, 1991.

61. DeGasperi, R., Thomas, L.J., Sugiyama, E., Chang, H.-M., Beck, P.J.,

Orlean, P., Albright, C., Waneck, G., Sambrook, J.F., Warren, C.D., and Yeh, E.T.H. Correction of a defect in mammalian GPI anchor biosynthesis by a transfected yeast gene. Science 250, 988, 1990.

62. Puoti, A., Desponds, C., Fankhauser, C., and Conzelmann, A.

Characterization of a glycophospholipid intermediate in the biosynthesis of glycosylphosphatidylinositol anchors accumulating in

Field and Menon 34

the Thy-1 negative lymphoma line SIA-b. J. Biol. Chem. 266, 21051, 1991.

63. Sugiyama, E., DeGasperi, R., Urakaze, M., Chang, H.-M., Thomas, L.J.,

Hyman, R., Warren, C.D., and Yeh, E.T.H. Identification of defects in glycosylphosphatidylinositol anchor biosynthesis in the Thy-1 expression mutants. J. Biol. Chem. 266, 12119, 1991.

64. Homans, S.W., Ferguson, M.A.J., Dwek, R.A., Rademacher, T.W.,

Anand, R., and Williams, A.F. Complete structure of the glycosylphosphatidylinositol membrane anchor of rat brain Thy-1 glycoprotein. Nature 333, 269, 1988.

65. Bangs, J.A., Andrews, N.W., Hart, G.W., and Englund, P.T.

Posttranslational modification and intracellular transport of a trypanosome variant surface glycoprotein. J. Cell Biol. 103, 255, 1986.

66. McConville, M.J., and Bacic, A. A family of glycoinositol

phospholipids from Leishmania major: isolation, characterization and antigenicity. J. Biol. Chem. 264, 757, 1989.

67. McConville, M.J., and Bacic, A. The glycoinositolphospholipid profiles

of two Leishmania major strains that differ in lipophosphoglycan expression. Mol. Biochem. Parasitol. 38, 57, 1990.

68. McConville, M.J., Homans, S.W., Thomas-Oates, J.E., Dell, A., and

Bacic, A. Structures of the glycoinositol phospholipids from Leishmania major. J. Biol. Chem. 265, 7385, 1990.

69. McConville, M.J., and Blackwell, J.M. Developmental changes in the

glycosylated phosphatidyl-inositols of Leishmania donovani. J. Biol. Chem. 266, 15170, 1991.

70. Turco, S.J., Wilkerson, M.A., and Clawson, D.R. Expression of an

unusual acidic glycoconjugate in Leishmania donovani. J. Biol. Chem. 259, 3883, 1984

71. Turco, S.J., Hull, S.R., Orlandi, P.A., Shepard, S.D., Homans, S.W.,

Dwek, R.A., and Rademacher, T.W. Structure of the major carbohydrate fragment of the Leishmania donovani lipophosphoglycan. Biochemistry 26, 6233, 1987.

72. Turco, S.J., Orlandi, P.A., Homans, S.W., Ferguson, M.A.J.,Dwek, R.A.,

and Rademacher, T.W. Structure of the phosphosaccharide-inositol

Field and Menon 35

core of the Leishmania donovani lipophosphoglycan. J. Biol. Chem. 264, 6711, 1989.

73. McConville, M.J., Thomas-Oates, J.E., Ferguson, M.A.J., and Homans

S.W. Structure of the lipophosphoglycan from Leishmania major. J. Biol. Chem. 265, 19611, 1990.

74. McNeely, T.B., and Turco, S.J. Requirement of lipophosphoglycan for

intracellular survivial of Leishmania donovani within human monocytes. J. Immunol. 144, 2745, 1990.

75. Turco, S.J. The leishmanial lipophosphoglycan: a multifunctional

molecule. Exp. Parasitol. 70, 241, 1990. 76. Field, M.C., Medina-Acosta, E., and Cross, G.A.M. Characterisation of a

glycosylphosphatidylinositol membrane protein anchor precursor in Leishmania mexicana. Mol. Biochem. Parasitol. 48, 227, 1991.

77. Previato, J.O., Gorin, P.A.J., Mazurek, M., Xavier. M.T., Fournet,

B.,Wieruszesk, J.M., and Mendonca-Previato, L.Primary structure of the oligosaccharide chain of lipopeptidophosphoglycan of epimastigote forms of Trypanosoma cruzi. J. Biol. Chem. 265, 2518, 1990.

78. Lederkremer, R.M., Lima, C., Ramirez, M.I., Ferguson, M.A.J., Homans,

S.W., and Thomas-Oates, J., Complete structure of the glycan of lipopeptidophosphoglycan from Trypanosoma cruzi epimastigotes. J. Biol. Chem. 266, 23670, 1991.

79. Mayor, S., Menon, A.K. Structural analysis of the glycosylinositol

phospholipid anchors of membrane proteins. Methods: A Companion to Methods in Enzymology 1, 297, 1990.

80. Doering, T.L., Raper, J., Buxbaum, L.U., Hart, G.W., and Englund, P.T.

Biosynthesis of glycosylphosphatidylinositol protein anchors. Methods: A Companion to Methods in Enzymology 1, 288, 1990.

81. Conzelmann, A., Fankhauser, C., and Desponds, C. Myoinositol gets

incorporated into numerous membrane glycoproteins of Saccharomyces cerevisiae: incorporation is dependent on phosphomannomutase [SEC53]. The EMBO J. 9, 653, 1990.

82. Orlean, P. Dolichol phosphate mannose synthase is required in vivo

for glycosyl phosphatidylinositol membrane anchoring, O

Field and Menon 36

mannosylation, and N glycosylation of protein in Saccharomyces cerevisiae. Mol. Cell Biol. 10, 5796, 1990.

83. Lisanti, M.P., Field, M.C., Caras, I.W., Menon, A.K., and Rodriguez-

Boulan, E. Mannosamine, a novel inhibitor of glycosyl-phosphatidylinositol incorporation into proteins. The EMBO J. 10, 1969, 1991.

84. Masterson, W.J., and Ferguson, M.A.J. Phenylmethanesulphonyl

fluoride inhibits GPI anchor biosynthesis in the African trypanosome. The EMBO J. 10, 2041, 1991.

85. Bordier, C., Phase separation of integral membrane proteins in Triton

X-114 solution. J. Biol. Chem. 256, 1604, 1981. 86. Pryde, J.G. Triton X-114: a detergent that has come in from the cold.

Trends in Biochem. Sci. 11, 160, 1986. 87. Roberts, W.L., and Rosenberry, T.L. Selective radiolabeling and

isolation of the hydrophobic membrane-binding domain of human erythrocyte acetylcholinesterase. Biochemistry 25, 3091, 1986.

87a. Huang, K-S., Li, S., and Low, M. Glycosylphosphatidylinositol specific

phospholipase D. Methods Enzymol. 197 567 (1991). 88. Field, M.C., and Menon, A.K. Biosynthesis of Glycosyl-

Phosphatidylinositol Membrane Protein Anchors in 'Lipid Modifications of Proteins - A Practical Approach' (Hooper, N., Turner, A.J. eds), IRL Press, Oxford, U.K. Chapter 7, 155, 1992.

89. Masterson, W.J., Doering, T.L., Hart, G.W., and Englund, P.T. A novel

pathway for glycan assembly: biosynthesis of the glycosyl-phosphatidylinositol anchor of the trypanosome variant surface glycoprotein. Cell 56, 793, 1989.

90. Mayor, S., Menon, A.K., Cross, G.A.M., Ferguson, M.A.J., Dwek, R.A.,

and Rademacher, T.W. Glycolipid precursors for the membrane anchor of Trypanosoma brucei variant surface glycoproteins: I. Can structure of the phosphatidylinositol-specific phospholipase C sensitive and resistant glycolipids. J. Biol. Chem. 265, 6164, 1990.

91. Menon, A.K., Schwarz, R.T., Mayor, S., and Cross, G.A.M. Cell-free

synthesis of glycosyl-phosphatidylinositol precursors for the glycolipid membrane anchor of Trypanosoma brucei variant surface

Field and Menon 37

glycoproteins: structural characterization of putative biosynthetic intermediates. J. Biol. Chem. 265, 9033, 1990.

92. Menon, A.K., Mayor, S., and Schwarz, R.T. Biosynthesis of glycosyl-

phosphatidylinositol lipids in Trypanosoma brucei: involvement of mannosylphosphoryldolichol as the mannose donor. The EMBO J. 9, 4249, 1990.

93. Bangs, J.A., Hereld, D., Krakow, J.L., Hart, G.W., and Englund, P.T.

Rapid processing of the carboxyl terminus of a trypanosome variant surface glycoprotein. Proc. Natl. Acad. Sci. USA 82, 3207, 1985.

94. Ferguson, M.A.J., Duszenko, M., Lamont, G.S., Overath, P., and Cross,

G.A.M. Biosynthesis of Trypanosoma brucei variant surface glycoproteins: N-glycosylation and addition of a phosphatidylinositol membrane anchor. J. Biol. Chem. 261, 356, 1986.

95. Takami, N., Ogata, S., Oda, K., Misumi, Y., and Ikehara, Y. Biosynthesis

of placental alkaline phosphatase and its post-translational modification by glycophospholipid for membrane-anchoring. J. Biol. Chem. 263, 3016, 1988.

96. Conzelmann, A., Spiazzi, A., and Bron, C. Glycolipid anchors are

attached to Thy-1 glycoprotein rapidly after translation. Biochem. J. 246, 605, 1987.

97. Takami, N., Misumi, Y., Kuroki, M., Matsuoka, Y., and Ikehara, Y.

Evidence for carboxyl-terminal processing and glycolipid-anchoring of human carcinoembryonic antigen. J. Biol. Chem. 263, 12716, 1988.

98. He, H.-T., Finne, J., and Goridis, C. Biosynthesis, membrane

association, and release of N-CAM-120, a phosphatidylinositol-linked form of the neural cell adhesion molecule. J. Cell Biol. 105, 2489, 1987.

99. Conzelmann, A., Riezman, H., Desponds, C., and Bron, C. A major

125-kd membrane glycoprotein of Saccharomyces cerevisiae is attached to the lipid bilayer through an inositol-containing phospholipid. The EMBO J. 7, 2233, 1988.

100. Medof, M.E., Walter, E.I., Roberts, W.L., Haas, R., and Rosenberry, T.L.

Decay accelerating factor of complement is anchored to cells by a C-terminal glycolipid. Biochemistry 25, 6740, 1986.

Field and Menon 38

101. Bailey, C.A., Gerber, L., Howard, A.D., and Udenfriend, S. Processing at the carboxyl terminus of nascent placental alkaline phosphatase in a cell-free system: evidence for specific cleavage of a signal peptide. Proc. Natl. Acad. Sci. USA 86, 22, 1989.

102. Kodukula, K., Micanovic, R., Gerber, L., Tamburrini, M., Brink, L., and

Udenfriend, S. Biosynthesis of phosphatidylinositol glycan-anchored membrane proteins: design of a simple protein substrate to characterize the enzyme that cleaves the COOH-terminal signal peptide. J. Biol. Chem. 266, 4464, 1991.

103. Kodukula, K., Cines, D., Amthauer, R., Gerber, L., and Udenfriend, S.

Biosynthesis of phosphatidylinositol-glycan (PI-G)-anchored membrane proteins in cell-free systems: cleavage of the nascent protein and addition of the PI-G moiety depend on the size of the COOH-terminal signal peptide. Proc. Natl. Acad. Sci. USA 89, 1350, 1992.

104. Krakow, J.L., Hereld, D., Bangs, J.D., Hart, G.W., and Englund, P.T.

Identification of a glycolipid precursor of the Trypanosoma brucei variant surface glycoprotein. J. Biol. Chem. 261, 12147, 1986.

105. Menon, A.K., Mayor, S., Ferguson, M.A.J., Duszenko, M., and Cross,

G.A.M. Candidate glycophospholipid precursor for the glycosylphosphatidylinositol membrane anchor of Trypanosoma brucei variant surface glycoproteins. J. Biol. Chem. 263, 1970, 1988.

106. Elbein, A.D. Inhibitors of the biosynthesis and processing of N-linked

oligosaccharide chains. Annu. Rev. Biochem. 56, 497, 1987. 107. Mayor, S., Menon, A.K., and Cross, G.A.M. Galactose-containing

glycosylphosphatidylinositols in Trypanosoma brucei. J. Biol. Chem. 267, 754, 1992.

108. Field, M.C., Menon, A.K., and Cross, G.A.M. Developmental variation

of glycosylphosphatidylinositol membrane anchors in Trypanosoma brucei: in vitro biosynthesis of intermediates in the construction of the GPI-anchor of the major procyclic surface glycoprotein. J. Biol. Chem. 267, 5324, 1992.

109. Menon, A.K., Schwarz, R.T., Mayor, S., and Cross, G.A.M. Glycolipid

precursor of Trypanosoma brucei variant surface glycoproteins: incorporation of radiolabeled mannose and myristic acid in a cell free system. Biochem. Soc. Trans. Lond. 16, 996, 1988.

Field and Menon 39

110. Doering, T.L., Masterson, W.J., Englund, P.T., and Hart, G.W. Biosynthesis of the glycosyl phosphatidylinositol membrane anchor of the trypanosome variant surface glycoprotein: origin of the non-acetylated glucosamine. J. Biol. Chem. 264, 11168, 1989.

111. Lennarz, W.J. Protein glycosylation in the endoplasmic reticulum:

current topological issues. Biochemistry 26, 7205, 1987. 112. Hirschberg, C.B. and Snider, M.D. Topography of glycosylation in the

rough endoplasmic reticulum and Golgi apparatus. Annu. Rev. Biochem. 56, 63-87.

112a. Abeijon, C., and Hirschberg, C.B. Topography of glycosylation

reactions in the rough endoplasmic reticulum. Trends in Biochem. Sci. 17, 32, 1992.

113. Stevens, V.L., and Raetz, C.R.H. Defective

glycosylphosphatidylinositol biosynthesis in extracts of three Thy-1 negative lymphoma cell mutants. J. Biol. Chem. 266, 10039, 1991.

114. Hirose, S., Ravi, L., Hazra, S.V., and Medof, M.E. Assembly and

deacetylation of N-acetylglucosaminyl-plasmanylinositol in normal and affected paroxysmal nocturnal haemoglobinuria cells. Proc. Natl. Acad. Sci. USA 88, 3762, 1991.

115. Hyman, R. Cell-surface-antigen mutants of haematopoietic cells: tools

to study differentiation, biosynthesis and function. Biochem. J. 225, 27, 1985.

116. Hyman, R. Somatic genetic analysis of the expression of cell surface

molecules. Trends in Genetics 4, 5, 1988 117. Chapman, A., Trowbridge, I.S., Hyman, R., and Kornfeld, S. Structure

of the lipid-linked oligosaccharides that accumulate in Class E Thy-1-negative mutant lymphomas. Cell 17, 509, 1979.

118. Chapman, A., Fujimoto, K., and Kornfeld, S. The primary

glycosylation defect in Class E Thy-1 negative mutant mouse lymphoma cells is an inability to synthesize dolichol-P-mannose. J. Biol. Chem. 255, 4441, 1980.

119. Conzelmann, A., Spiazzi, A., Hyman, R., and Bron, C. Anchoring of

membrane proteins via phosphatidylinositol is deficient in two classes of Thy-1 negative mutant lymphoma cells. The EMBO J. 5, 3291, 1986.

Field and Menon 40

120. Fatemi, S.H., and Tartakoff, A.M. Hydrophilic anchor-deficient Thy-1

is secreted by a class E mutant T lymphoma. Cell 46, 653, 1986. 121. Fatemi, S.H., Haas, R., Jentoft, N., Rosenberry, T.L., and Tartakoff, A.M.

The glycophospholipid anchor of Thy-1: biosynthetic labelling experiments with wild-type and Class E Thy-1 negative lymphomas. J. Biol. Chem. 262, 4728, 1987.

122. Fatemi, S.H., and Tartakoff, A.M. The phenotype of five classes of T

lymphoma mutants: defective glycophospholipid anchoring, rapid degradation, and secretion of Thy-1 glycoprotein. J. Biol. Chem. 263, 1288, 1988.

123. Yeh, E.T.H., Reiser, H., Bamezai, A., and Rock, K. TAP transcription

and phosphatidylinositol linkage mutants are defective in activation through the T-cell receptor. Cell 52, 665, 1988.

124. Schwarz, R.T., Mayor, S., Menon, A.K., and Cross, G.A.M. Biosynthesis

of the glycolipid membrane anchor of Trypanosoma brucei variant surface glycoprotein: involvement of Dol-P-Man. Biochem. Soc. Trans. Lond. 17, 746, 1989.

125. Takami, N., Oda, K., and Ikehara, Y. Aberrant processing of alkaline

phosphatase precursor caused by blocking the synthesis of glycosylphosphatidylinositol. J. Biol. Chem. 267, In press, 1992.

126. Menon, A.K., and Stevens, V.L. Phosphatidylethanolamine is the

donor of the ethanolamine residue linking a glycosylphosphatidylinositol anchor to protein. J. Biol. Chem. 267, -- ,1992.

127. Menon, A.K., Eppinger, M., Mayor, S., and Schwarz, R.T. Biosynthesis

of glycosylphosphatidylinositol anchors in Trypanosoma brucei: the terminal phosphoethanolamine group is derived from phosphatidylethanolamine. Submitted.

128. Masterson, W.J., Raper, J., Doering, T.L., Hart, G.W., and Englund, P.T.

Fatty acid re-modelling: a novel reaction sequence in the biosynthesis of trypanosome glycosyl phosphatidylinositol membrane anchors. Cell 62, 73, 1990.

129. Bütikofer, P., Kuypers, F.A., Shackleton, C., Brodbeck, U., and Stieger,

S. Molecular species analysis of the glycosylphosphatidylinsotiol

Field and Menon 41

anchor of Torpedo marmorata acetylcholinesterase. J. Biol. Chem. 265, 18983, 1990.

130. Clayton, C.E., and Mowatt, M.R. The procyclic acidic repetitive

proteins of Trypanosoma brucei: purification and post-translational modification. J. Biol. Chem. 264, 15088, 1989.

131. Costello, L.C., and Orlean, P. Inositol acylation of a potential glycosyl

phosphoinositol anchor precursor from yeast requires acyl co-enzyme A. J. Biol. Chem. 267, ----, 1992.

132. Lamanna, C., and Jones, L., Lethality for mice of vegetative and spore

forms of Bacillus cereus and Bacillus cereus-like insect pathogens injected intraperitoneally and subcutaneously. J. Bacteriol. 85, 532, 1963.

133. Lemansky, P., Gupta, D.K., Meyale, S., Tucker, G., and Tartakoff, A.M.

Atypical mannolipids characterize Thy-1-negative lymphoma mutants. Mol. Cell Biol. 11, 3879, 1991.

134. Urakaze, M., Kamitani, T., DeGasperi, R., Sugiyama, E., Chang, H.-M.,

Warren, C.D., and Yeh, E.T.H. Identification of a missing link in glycosylphosphatidylinositol anchor biosynthesis in mammalian cells. J. Biol. Chem. 267, In press, 1992.

135. Toutant, Richards, Krall, and Rosenberry. Molecular forms of

acetylcholinesterase in two sublines of human erythroleukemia K562 cells: sensitivity to PI-PLC and biosynthesis. Eur. J. Biochem. 180, 503, 1989.

136. Singh, N., Singleton, D., and Tartakoff, A.M. Anchoring and

degradation of glycolipid-anchored membrane proteins by L929 versus by LM-TK- mouse fibroblasts: implications for anchor biosynthesis. Mol. Cell Biol. 11, 2362, 1991.

137. Field, M.C. Inositol Acylation of Glycosyl-phosphatidylinositol

Glycoconjugate J. 9, -- , 1992 138. Bell, R.M., Ballas, L.M., and Coleman, R.A. Lipid topogenesis. J. Lipid

Res. 22, 391, 1981. 139. Haselbeck, A., and Tanner, W. Dolichyl phosphate-mediated

mannosyl transfer through liposomal membranes. Proc. Natl. Acad. Sci. USA 79, 1520, 1982.

Field and Menon 42

140. Higgins, J.A., and Pigott, C.A. Asymmetric distribution of phosphatidylethanolamine in the endoplasmic reticulum demonstrated using trinitrobenzene-sulphonic acid as a probe. Biochim. Biophys. Acta 693, 151, 1982.

141. Mayor, S. Biosynthesis of glycosyl-phosphatidylinositol membrane

protein anchors: study of Trypanosoma brucei variant surface glycoprotein GPI-anchor. Ph. D. Thesis, The Rockefeller University, 1991.

142. Datema, R., Schwarz, R.T., and Jankowski, A.W. Fluoroglucose-

inhibition of protein glycosylation in vivo: inhibition of mannose and glucose incorporation into lipid-linked oligosaccharides. Eur. J. Biochem. 109, 331, 1980a.

143. Datema, R., Schwarz, R.T., and Winkler, J. Glycosylation of influenza

virus proteins in the presence of fluoroglucose occurs via a different pathway. Eur. J. Biochem. 110, 355, 1980b.

144. Schwarz, R.T., and Datema, R. The lipid pathway of protein

glycosylation and its inhibitors: the biological significance of protein-bound carbohydrates. Adv. Carbohyd. Chem. Biochem. 40, 287, 1982.

145. Banerjee, D.K., Scher, M.G., and Waechter, C.J. Amphomycin: effect of

the lipopeptide antibiotic on the glycosylation and extraction of dolichyl monophosphate in calf brain membranes. Biochemistry 20, 1561, 1981.

146. Banerjee, D.K. Amphomycin inhibits mannosylphosphoryl-dolichol

synthesis by forming a complex with dolichylmonophosphate. J. Biol. Chem. 264, 2024, 1989.

147. Pan, Y.T., and Elbein, A.D. The effect of mannosamine on the

formation of lipid-linked oligosaccharides and glycoproteins in canine kidney cells. Arch. Biochem. Biophys. 242, 447, 1985.

148. Hirose, S., Mohney, R.P., Mutka, S.C., Ravi, L., Singleton, D.R., Perry, G.,

Tartakoff, A.M., and Medof, M.E. Derivation and characterization of glycoinositol-phospholipid anchor-defective human K562 cell clones. J. Biol. Chem. 267, 5272, 1992.

149. Hollander, N., Selvaraj, P., and Springer, T.A. Biosynthesis and

function of LFA-3 in human mutant cells deficient in

Field and Menon 43

phosphatidylinositol-anchored proteins. J. Immunol. 141, 4283, 1988.

150. Teng, M.-H., Hedayati, S., Alexander, A.A., Barkin, R., Basch, R.S., and

Buxbaum, J.N. A Thy-1 negative lymphoma cell variant defective in the formation of glycosyl-phosphatidylinositol membrane protein anchors. Molec. Immunol. 26, 391, 1989.

151. Stroynowski, I., Soloski, M., Low, M.G., and Hood, L. A single gene

encodes soluble and membrane-bound forms of the major histocompatibility Qa-2 antigen: anchoring of the product by a phospholipid tail. Cell 50, 759, 1987.

152. Waneck, G.L., Sherman, D.H., Kincade, P.W., Low, M.G., and Flavell,

R.A. Molecular mapping of signals in the Qa-2 antigen required for attachment of the phosphatidylinositol membrane anchor. Proc. Natl. Acad. Sci. USA 85, 577, 1988.

153. Singh, N., and Tartakoff, A.M. Two different mutants blocked in the

synthesis of dolichol-phosphoryl-mannose do not add glycophospholipid anchors to membrane proteins: quantitative correction of the phenotype of a CHO cell mutant with tunicamycin. Mol. Cell Biol. 11, 391, 1991.

154. Conzelmann, A., Spiazzi, A., Bron, C., and Hyman, R. No glycolipid

anchors are added to Thy-1 glycoprotein in Thy-1-negative mutant thymoma cells of four different complementation classes. Mol. Cell Biol. 8, 674, 1988.

155. Stevens, V.L., and Raetz, C.R.H. Class F Thy-1-negative murine

lymphoma cells are deficient in ether lipid biosynthesis. J. Biol. Chem. 265, 15653, 1990.

156. Gupta, D., Tartakoff, A., and Tisdale, E. Metabolic correction of

defects in the lipid anchoring of Thy-1 in lymphoma mutants. Science 242, 1446, 1988.

157. Nicholson-Weller, A., March, J.P., Rosenfeld, S.I., and Austen, K.F.

Affected erythrocytes of patients with paroxysmal nocturnal hemoglobinuria are deficient in the complement regulatory protein, decay accelerating factor. Proc. Natl. Acad. Sci. USA 80, 5066, 1983.

158. Pangburn, M.K., Schreiber, R.D., and Müller-Eberhard, H.J. Deficiency

of an erythrocyte membrane protein with complement regulatory

Field and Menon 44

activity in paroxysmal nocturnal hemoglobinuria. Proc. Natl. Acad. Sci. USA 80, 5430, 1983.

159. Zalman, L.S., Wood, L.M., Frank, M.M., and Müller-Eberhard, H.J.

Deficiency in the homologous restriction factor in paroxysmal nocturnal hemoglobinuria. J. Exp. Med. 165, 572, 1987.

160. Selvaraj, P., Dustin, M.L., Silber, R., Low, M.G., and Springer, T.A.

Deficiency of lymphocyte function-associated antigen 3 [LFA-3] in paroxysmal nocturnal hemoglobinuria. J. Exp. Med. 166, 1011, 1987.

161. Burroughs, S.F., Devine, D.V., Browne, G., and Kaplan, M.E. The

population of paroxysmal nocturnal hemoglobinuria neutrophils deficient in decay-accelerating factor is also deficient in alkaline phosphatase. Blood 71, 1086, 1988.

162. Simmons, D.L., Tan, S., Tenen, D.G., Nicholson-Weller, A., and Seed, B.

Monocyte antigen CD14 is a phospholipid-anchored membrane protein. Blood 73, 284, 1989.

163. Selvaraj, P., Rosse, W.F., Silber, R., and Springer, T.A. The major Fc

receptor in blood has a phosphatidylinositol anchor and is deficient in paroxysmal nocturnal hemoglobinuria. Nature 333, 565, 1988.

164. Stafford, H.A., Tykocinski, M.L., Lublin, D.M., Holers, V.M., Rosse, W.F.,

Atkinson, J.P., and Medof, M.E. Normal polymorphic varaitions and transcription of the decay accelerating factor gene in paroxysmal nocturnal hemoglobinuria cells. Proc. Natl. Acad. Sci. USA 85, 880, 1988.

165. Carothers, D.J., Hazra, S.V., Andreson, S.W., and Medof, M.E.

Synthesis of aberrant decay-accelerating factor proteins by affected paroxysmal nocturnal hemoglobinuria leukocytes. J. Clin. Invest. 85, 47, 1990.

165a. Schubert, J., Uciechowski, P., Delany, P., Tischler, H.-J., Kolanus, W.,

and Schmidt, R.E. The PIG-anchoring defect in NK lymphocytes of PNH patients. Blood 76, 1181, 1990.

166. Boothroyd, J.C., Paynter, C.A., Cross, G.A.M., Bernards, A., and Borst,

P. Variant surface glycoproteins of Trypanosoma brucei are synthesized with cleavable hydrophobic sequences at the carboxy and amino termini. Nucleic Acids Res. 9, 4735, 1981.

Field and Menon 45

167. Caras, I.W.,Weddell, G.N., Davitz, M.A., Nussenzweig, V., and Martin, D.W. Signal for attachment of a phospholipid membrane anchor in decay accelerating factor. Science 238, 1280, 1987.

167a. Caras, I.W.,Weddell, G.N., and Williams, S.R. Analysis of the signal for

attachment of a glycophospholipid membrane anchor. J. Cell Biol. 108, 1387, 1989.

168. Berger, J., Howard, A.D., Gerber, L., Hauber, J., Cullen, B.R., and

Udenfriend, S. Expression of active, membrane-bound human placental alkaline phosphatase by transfected simian cells. Proc. Natl. Acad. Sci. USA 84, 4885, 1987.

169. Berger, J., Howard, A.D., Brink, L., Gerber, L., Hauber, J., Cullen, B.R.,

and Udenfriend, S. COOH-terminal requirements for the correct processing of a phosphatidylinositol-glycan anchored membrane protein. J. Biol. Chem. 263, 10016, 1988.

170. Berger, J., Micanovic, R., Greenspan, R.J., and Udenfriend, S.

Conversion of placental alkaline phosphatse from a phosphatidylinositol-glycan-anchored protein to an integral transmembrane protein. Proc. Natl. Acad. Sci. USA 86, 1457, 1989.

171. Tykocinski, M.L., Shu, H.-K., Ayers, D.J., Walter, E.I., Getty, R.R.,

Groger, R.K., Hauer, C.A., and Medof, M.E. Glycolipid re-anchoring of T-lymphocyte surface antigen CD8 using the 3' end sequence of decay accelerating factor's mRNA. Proc. Natl. Acad. Sci. USA 85, 3555, 1988.

172. Caras, I.W., and Weddell, G.N. Signal peptide for protein secretion

directing glycophospholipid membrane anchor attachment. Science 243, 1196, 1989.

173. Udenfriend, S., Micanovic, R., and Kodukula, K. Structural

requirements of a nascent protein for processing to a PI-G anchored form: studies in intact cells and cell-free systems. Cell Biol. Int. Rep. 15, 739, 1991.

174. Caras, I.W. An internally positioned signal can direct attachment of a

glycophospholipid membrane anchor. J. Cell Biol. 113, 77, 1991. 175. Micanovic, R., Gerber, L.D., Berger, J., Kodukula, K., and Udenfriend,

S. Selectivity of the cleavage/attachment site of phosphatidylinositol-glycan-anchored membrane proteins determined by site-specific

Field and Menon 46

mutagenesis at Asp-484 of placental alkaline phosphatase. Proc. Natl. Acad. Sci. USA 87, 157, 1990.

176. Moran, P., Raab, H., Kohr, W.J., and Caras, I.W. Glycophospholipid

membrane anchor attachment: molecular analysis of the cleavage/attachment site. J. Biol. Chem. 266, 1250, 1991.

177. Gerber, L.D., Kodukula, K., and Udenfriend, S. Phosphatidylinositol-

glycan (PI-G) anchored membrane proteins: amino acid requirements adjacent to the site of cleavage and PI-G attachment in the COOH-terminal signal peptide. J. Biol. Chem., 267, ---, 1992.

178. Moran, P., and Caras, I.W. A nonfunctional sequence converted to a

signal for glycosylphosphatidylinositol membrane anchor attachment. J. Cell Biol. 115, 329, 1991.

179. Dustin, M.L., Selvaraj, P., Mattaliano, R.J., and Springer, T.A.

Anchoring mechanisms for LFA-3 cell adhesion glycoprotein at membrane surface. Nature 329, 846, 1987.

180. Barthels, D., Santoni, M.-J., Wille, W., Ruppert, C., Chaix, J.-C., Hirsch,

M.R., Fontecilla-Camps, J.C., and Goridis, C. Isolation and nucleotide sequence of mouse NCAM cDNA that codes for a Mr 79,000 polypeptide without a membrane-spanning region. The EMBO J. 6, 907, 1987.

181. Owens, G.C., Edelman, G.M., and Cunningham, B.A. Organization of

the neural cell adhesion molecule (N-CAM) gene: alternative exon usage as the basis for different membrane-associated domains. Proc. Natl. Acad. Sci. USA 84, 294, 1987.

182. Hemperley, J.J., Edelman, G.M., and Cunningham, B.A. cDNA clones

of neural cell adhesion molecule [N-CAM] lacking a membrane-spanning region consistent with evidence for membrane attachment via a phosphatidylinositol intermediate. Proc. Natl. Acad. Sci. USA 83, 9822, 1986.

183. Hemperley, J.J., Murray, B.A., Edelman, G.M., and Cunningham, B.A.

Sequence of a cDNA clone encoding the polysialic acid-rich and cytoplasmic domains of the neural cell adhesion molecule N-CAM. Proc. Natl. Acad. Sci. USA 83, 3037, 1986.

184. Ravetch, J.V., and Perussia, B. Alternative membrane forms of Fcγ RIII

[CD16] on human natural killer cells and neutrophils: cell type-

Field and Menon 47

specific expression of two genes that differ in single nucleotide substitutions. J. Exp. Med. 170, 481, 1989.

185. Mayor, S., Menon, A.K., and Cross, G.A.M. Transfer of glycosyl-

phosphatidylinositol membrane anchors to polypeptide acceptors in a cell-free system. J. Cell Biol. 114, 61, 1991.

186. Thomas, L.J., Urakaze, M., DeGasperi, R., Kamitani, T., Sugiyama, E.,

Chang, H.-M., Warren, C.D., and Yeh, E.T.H. Differential expression of glycosylphosphatidylinositol-anchored proteins in a murine T-cell hybridoma mutant producing limiting amounts of the glycolipid core: implications for paroxysmal nocturnal hemoglobinuria. J. Clin. Invest., In press,1992.

187. Edberg, J.C., Salmon, J.E., Whitlow, M., and Kimberly, R.P. Preferential

expression of human FcgRIIIPMN (CD16) in paroxysmal nocturnal hemoglobinuria: discordant expression of glycosyl phosphatidylinositol-linked proteins. J. Clin. Invest. 87, 58, 1991.

188. Fasel, N., Rousseaux, M., Schaerer, E., Medof, M.E., Tykocinski, M.L.,

and Bron, C. In vitro attachment of glycosyl-inositol phospholipid anchor structures to mouse Thy-1 antigen and human decay-accelerating factor. Proc. Natl. Acad. Sci. USA 86, 6858, 1989.

189. Micanovic, R., Kodukula, K., Gerber, L.D., and Udenfriend, S.

Selectivity at the cleavage/attachment site of phosphatidylinositol-glycan anchored protein is enzymatically determined. Proc. Natl. Acad. Sci. USA 87, 7939, 1990.

190. Strominger, J.L., Izaki, K., Matsuhashi, M., and Tipper, D.J.

Peptidoglycan transpeptidase and D-alanine carboxypeptidase: penicillin-sensitive enzymatic reactions. Fed. Proc. 26, 9, 1967.

191. Tipper, D.J., and Strominger, J.L. Mechanism of action of penicillins:

a proposal based on their structural similarity to acyl-D-alanyl-D-alanine. Proc. Natl. Acad. Sci. USA 54, 1133, 1965.

192. Wise, E.M., and Park, J.T. Penicillin: its basic site of action as an

inhibitor of a peptide cross-linking reaction in cell wall mucopeptide synthesis. Proc. Natl. Acad. Sci. USA 54, 75, 1965.

193. Berne, P., Schmitter, J., and Blanquet, S. Peptide and carboxy-

terminal labeling through carboxypeptidase Y-catalyzed transpeptidation. J. Biol. Chem. 265, 19551, 1990.

Field and Menon 48

195. Walsh, C. Enzymatic reaction mechanisms, W.H. Freeman & Co., San Francisco, CA, 1979.

196. Bowles, D.J., and Pappin, D.J. Traffic and assembly of concanavalin A.

Trends in Biochem. Sci. 13, 60, 1988. 197. Hendrix, R.W. Protein carpentry. Current Biology 1, 71, 1991. 198. Kane, P.M., Yamishiro, C.T., Wolczyk, D.F., Neff, N., Goebl, M., and

Stevens, T.H. Protein splicing converts the yeast TFP1 gene product to the 69kD subunit of the vacuolar H+-adenosine triphosphatase. Science 250, 651, 1990.

199. Tank, D.W., Wu, W.-S., and Webb, W.W. Enhanced molecular

diffusibility in muscle membrane blebs: release of lateral constraints. J. Cell Biol. 92, 207, 1982.

200. Wier, M., and Edidin, M. Constraint of the translational diffusion of a

membrane glycoprotein by its external domains. Science 242, 412, 1988.

201. Zhang, F., Crise, B., Su, B., Hou, Y., Rose, J.K., Bothwell, A., and

Jacobson, K. Lateral diffusion of membrane-spanning and glycosylphosphatidylinositol-linked proteins: toward establishing rules governing the lateral mobility of membrane proteins. J. Cell Biol. 115, 75, 1991.

202. Rothberg, K.G., Ying, Y., Kolhouse, J.F., Kamen, B.A., and Anderson,

R.G.W. The glycophospholipid-linked folate receptor internalizes folate without entering the clathrin-coated pit endocytic pathway. J. Cell Biol. 110, 637, 1990a.

203. Rothberg, K.G., Ying, Y., Kamen, B.A., and Anderson, R.G.W.

Cholesterol controls the clustering of the glycophospholipid-anchored membrane receptor for 5-methyltetrahydrofolate. J. Cell Biol. 111, 2931, 1990b.

204. Edidin, M., and Stroynowski, I. Differences between the lateral

organization of conventional and inositol phospholipid-anchored membrane proteins: a further definition of micrometer scale membrane domains. J. Cell Biol. 112, 1143, 1991.

205. Edidin, M., Kuo, S.C., and Sheetz, M.P. Lateral movements of

membrane glycoproteins restricted by dynamic cytoplasmic barriers. Science 254, 1379, 1991.

Field and Menon 49

206. Hopkins, C. Polarity signals. Cell 66, 827, 1991. 207. Simons, K., and Wandinger-Ness, A. Polarized sorting in epithelia.

Cell 62, 207, 1990. 208. Brown, D.A., Crise, B., and Rose, J.K. Mechansim of membrane

anchoring affects polarized expression of two proteins in MDCK cells. Science 245, 1499, 1989.

209. Hooper, N.M., and Turner, A.J. Ectoenzymes of the kidney microvillar

membrane. FEBS Lett. 229, 340, 1988. 210. Takesue, Y., Yokota, K., Nishi, Y., Taguchi, R., and Ikezawa, H.

Solubilization of trehalase from rabbit renal and intestinal brush-border membranes by a phosphatidylinositol-specific phospholipase C. FEBS Lett. 201, 5, 1986.

211. Lisanti, M.P., Sargiacomo, M., Graeve, L., Saltiel, A.R., and Rodriguez-

Boulan, E. Polarized apical distribution of glycosyl-phosphatidylinositol anchored proteins in a renal epithelial cell line. Proc. Natl. Acad. Sci. USA 85, 9557, 1988.

212. Lisanti, M.P., and Rodriguez-Boulan, E. Glycophospholipid membrane

anchoring provides clues to the mechanism of protein sorting in polarized epithelial cells. Trends in Biochem. Sci. 15, 113, 1990.

213. Lisanti, M.P., and Rodriguez-Boulan, E. Polarized sorting of GPI-linked

proteins in epithelia and membrane microdomains. Cell Biol. Int. Rep. 15, 1023, 1991.

214. Wilson, J.M., Fasel, N., and Kraehenbuhl, J.-P. Polarity of endogenous

and exogenous GPI-anchored membrane proteins in MDCK cells. J. Cell Sci. 96, 143, 1990.

215. Lisanti, M.P., Caras, I.W., Davitz, M.A., and Rodriguez-Boulan, E. A

glycophospholipid membrane anchor acts as an apical targeting signal in polarized epithelial cells. J. Cell Biol. 109, 2145, 1989.

216. Dotti, C.G., Parton, R.G., and Simons, K. Polarized sorting of glypiated

proteins in hippocampal neurons. Nature 349, 158, 1991. 217. Brown, D.A., and Rose, J.K. Sorting of GPI-anchored proteins to

glycolipid-enriched membrane subdomains during transport to the cell surface. Cell 68, 533, 1992.

Field and Menon 50

218. Bretscher, M.S., Thomson, J.N., and Pearse, B.M.F. Coated pits act as

molecular filters. Proc. Natl. Acad. Sci. USA 77, 4156, 1980. 219. Lemansky, P., Fatemi, S.H., Gorican, B., Meyale, S., Rossero, R., and

Tartakoff, A.M. Dynamics and longevity of the glycolipid-anchored membrane protein, Thy-1. J. Cell Biol. 110, 1525, 1990.

220. Matsuura, S., Eto, S., Kato, K., and Tashiro, Y. Ferritin

immunoelectron microscopic localization of 5’-nucleotidase on rat liver cell surface. J. Cell Biol. 99, 166, 1984.

221. Rothberg, K.G., Heuser, J.E., Donzell, W.C., Ying, Y.-S., Glenney, J.R.,

and Anderson, R.G.W. Caveolin, a protein component of caveolae membrane coats. Cell 68, 673, 1992.

222. Anderson, R.G.W., Kamen, B.A., Rothberg, K.G., and Lacey, S.W.

Potocytosis: sequestration and transport of small molecules by caveolae. Science 255, 410, 1992.

223. Robinson, P.J. Phosphatidylinositol membrane anchors and T-cell

activation. Immunol. Today 12, 35, 1991. 224. Kroczek, R.A., Gunter, K.C., Germain, R.N., and Shevach, E.M. Thy-1

functions as a signal transduction molecule in T lymphocytes and transfected B lymphocytes. Nature 322, 181, 1986.

225. Robinson, P.J., Millrain, M., Antoniou, J., Simpson, E., and Mellor, A.L.

A glycophospholipid anchor is required for Qa-2-mediated T cell activation. Nature 342, 85, 1989.

226. Su, B., Waneck, G.L., Flavell, R.A., and Bothwell, A.L.M. The glycosyl

phosphatidylinositol anchor is critical for Ly-6A/E-mediated T cell activation. J. Cell Biol. 112, 377, 1991.

227. Sussman, J.J., Saito, T., Shevach, E., Germain, R.N., and Ashwell, J.D.

Thy-1- and Ly-6-mediated lymphokine production and growth inhibition of a T cell hybridoma require co-expression of the T cell antigen receptor complex. J. Immunol. 140, 2520, 1988.

228. Thomas, L.J., DeGasperi, R., Sugiyama, E., Chang, H.-M., Beck, P.J.,

Orlean, P., Urakaze, M., Kamitani, T., Sambrook, J.F., Warren, C.D., and Yeh, E.T.H. Functional analysis of T-cell mutants defective in the biosynthesis of glycosylphosphatidylinositol anchor: relative

Field and Menon 51

importance of glycosylphosphatidylinositol anchor versus N-linked glycosylation in T-cell activation. J. Biol. Chem. 266, 23175, 1991.

229. Presky, D.H., Low, M.G., and Shevach, E.M. Role of

phosphatidylinositol-anchored proteins in T cell activation. J. Immunol. 144, 860, 1990.

230. Stiernberg, J., Low, M.G., Flaherty, L., and Kincade, P.W. Removal of

lymphocyte surface molecules with phosphatidylinositol-specific phospholipase C: effects on mitogen responses and evidence that ThB and certain Qa antigens are membrane-anchored via phosphatidylinositol. J. Immunol. 38, 3877, 1987.

231. Einhorn, G., Qin, L., and Soloski, M.J. Biosynthesis of

glycophospholipid bound and secreted murine Class I Qa-2 polypeptides. Molec. Immunol. 28, 1299, 1991.

232. Soloski, M.J., Lattimore, A., Hereld, D., Krakow, J.L., Low, M.G., and

Einhorn, G. Further characterization of the membrane anchor found on the tissue-specific Class I molecule Qa2. J. Immunol. 140, 3858, 1988.

233. Ting, A.E., and Pagano, R.E. Detection of a phosphatidylinsotiol-

specific phospholipase C at the surface of Swiss 3T3 cells and its potential role in the the regulation of cell growth. J. Biol. Chem. 265, 5337, 1990.

234. Romero, G., Luttrell, L., Rogol, A., Zeller, K., Hewlett, E., and Larner, J.

Phosphatidylinositol-glycan anchors of membrane proteins: potential precursors of insulin mediators. Science 240, 509, 1988.

235. Hoessli, D., and Rungger-Brändle, E. Association of specific cell-

surface glycoproteins with a Triton-X-100-resistant complex of plasma membrane proteins isolated from T-lymphoma cells (P1798). Exp. Cell Res. 156, 239, 1985.

236. Stefanová, I., Horejsí, V., Ansotegui, I.J., Knapp, W., and Stockinger, H.

GPI-anchored cell-surface molecules complexed to protein tyrosine kinases. Science 254, 1016, 1991.

237. Schenkman, S., Yoshida, N., and Cardoso de Almeida, M.L.

Glycophosphatidylinositol-anchored proteins in metacyclic trypomastigotes of Trypanosoma cruzi. Mol. Biochem. Parasitol. 29, 141, 1988.

Field and Menon 52

238. Hernandez-Munain, C., Fernandez, M.A., Alcina, A., and Fresno, M. Characterization of a glycosyl-phosphatidylinositol-anchored membrane protein from Trypanosoma cruzi'. Infec. Immun. 59, 1409, 1991

239. Fouts, D.L., Ruef, B.J., Ridley, P.T., Wrightsman, R.A., Peterson, D.S.,

and Manning, J.E. Nucleotide sequence and transcription of a trypomastigote surface antigen gene of Trypanosoma cruzi'. Mol. Biochem. Parasitol. 46, 189, 1991.

240. Takle, G.B., and Cross, G.A.M. An 85kDal surface antigen gene family

of Trypanosoma cruzi encodes polypeptides homologous to bacterial neuraminidases. Mol. Biochem. Parasitol. 48, 57, 1991.

241. Pereira, M.E.A., Mejia, J., Ortega-Barria, E., Matzilevich, D., and Prioli,

R.P. The Trypanosoma cruzi neuraminidase contains sequences similar to bacterial neuraminidases, YWTD repeats of the low density lipoprotein receptor, and type III modules of fibronectin. J. Exp. Med. 174, 179, 1991

242. Kahn, S., Colbert, T.G., Wallace, J.C., Hoagland, N.A., and Eisen, H. The

major 85kDal surface antigen of the mammalian-stage forms of Trypanosoam cruzi is a family of sialidases. Proc. Natl. Acad. Sci. USA 88, 4481, 1991.

243. Voorhis, W.C.V., Schlekewy, L., and Trong, H.L. Molecular mimicry by

Trypanosoma cruzi: the Fl-160 epitope that mimics mammalian nerve can be mapped to a 12-amino acid peptide. Proc. Natl. Acad. Sci. USA 88, 5993, 1991.

244. Pollevick, G.D., Affranchino, J.L., Frasch, A.C.C., and Sanchez, D.O. The

complete sequence of a shed acute-phase antigen of Trypanosoma cruzi. Mol. Biochem. Parasitol. 47, 247, 1991.

245. Rosenberg, I., Prioli, R.P., Ortega-Barria, E., and Pereira, M.E.A. Stage-

specific phospholipase C-mediated release of Trypanosoma cruzi neuraminidase. Molec. Biochem. Parasitol. 46, 303, 1991.

246. Engstler, M., Reuter, G., and Schauer, R. Partial purification and

characterization of a sialidase from T. brucei. Glycoconjugate J. 8, 269, 1991.

247. Schell, D., Evers, R., Preis, D., Ziegelbauer, K., Kiefer, H., Lottspeich, F.,

Cornelssen, A.W.C.A., and Overath, P. A transferrin-binding protein of Trypanosoma brucei is encoded by one of the genes in the variant

Field and Menon 53

surface glycoprotein gene expression site. The EMBO J. 10, 1061, 1991.

248. Rivas, L., Kahl, L., Manson, K., and McMahon-Pratt, D. Biochemical

chracterisation of the protective membrane glycoprotein GP42/M-2 of Leishmania amazonensis. Mol. Biochem. Parasitol. 47, 235, 1991.

249. Medina-Acosta, E., Karess, R.E., Schwartz, H., and Russell, D.G. The

promastigote surface protease (gp63) of Leishmania is expressed but differentially processed and localized in the amastigote stage. Mol. Biol. Parasitol. 37, 263, 1989.

250. Murray, P.J., Spithill, T.W., and Handman, E. The PSA-2 glycoprotein

of L. major is a glycosylphosphatidylinositol anchored surface antigen. J. Immunol. 143, 4221, 1989

251. Das, S., Traynor-Kaplan, A., Reiner, D.S., Meng, T.-C., and Gillin, F.D.

A surface antigen of Giardia lamblia with a glycosylphosphatidylinositol anchor. J. Biol. Chem. 266, 21318, 1991.

252. Camerini, D., James, S.P., Stamenkovic, I., and Seed, B. Leu8/TQ1 is

the human equivalent of the Mel-14 lymph node homing receptor. Nature 342, 78, 1989.

253. Symington, B.E., and Symington, F.W. Evidence for

phosphatidylinositol-linked forms of human and avian fibronectin receptors. J. Biol. Chem. 265, 14069, 1990.

254. Karagogeos, D., Morton, S.B., Casano, F., Dodd, J., and Jessell, T.M.

Developmental expression of the axonal glycoprotein TAG-1; differential regulation by central and peripheral neurones in vitro. Development 112, 51, 1991.

254a. Furley, A.J., Morton, S.B., Manalo, D., Karargogeous, D., Dodd, J., and

Jessell, T.M. The axonal glycoprotein TAG-1 is an immunolglobulin superfamily member with neurite outgrowth-promoting activity. Cell 61, 157, 1990.

255. Schofield, P.R., McFarland, K.C., Hayflick, J.S., Wilcox, J.N., Cho, T.,

Roy, S., Lee, N.M., Loh, H.H., and Seeburg, P.H. Molecular characterisation of a new immunoglobulin superfamily protein with potential roles in opioid binding and cell contact. The EMBO J. 8, 489, 1989.

Field and Menon 54

256. Hortsch, M., and Goodman, C.S. Drosophila fasciculin I, a neural cell adhesion molecule, has a phosphatidylinositol lipid membrane anchor that is developmentally regulated. J. Biol. Chem. 265, 15104, 1990.

257. Krantz, D.E., and Zipursky, S.L., Drosophila chaoptin, a member of the

leucine-rich repeat family, is a photoreceptor cell-specific adhesion molecule. The EMBO J. 9, 175, 1990.

258. Lipke, P.N., Wojciechowicz, D., and Kurjan, J. AGα1 is the structural

gene for the Saccharomyces cerevisiae α-agglutinin, a cell surface glycoprotein involved in cell-cell interactions during mating. Mol. Cell. Biol. 9, 3155, 1989.

259. Roy, A., Lu, C.F., Marykwas, D.L., Lipke, P.N., and Kurjan, J. The AGA1

product is involved in cell-surface attachemnet of the Saccharomyces cerevisiae cell-adhesion glycoprotein A-agglutinin. Mol. Cell. Biol. 11, 4196, 1991

260. Ploug, M., Ronne, E., Behrendt, N., Jensen, A.L., Blasi, F., and Dano,

K.Cellular receptor for urokinase plasminogen-activator: carboxyl-terminal processing and membrane anchoring by glycosyl-phosphatidylinositol. J. Biol. Chem. 266, 1926, 1991.

261. Skubitz, K.M., Stroncek, D.F and Sun, B Neutrophil-specific antigen

NB-1 is anchored via a glycosyl-phosphatidylinositol linkage. J. Leuk. Biol. 49, 163, 1991.

262. Cashman, N.R., Loertscher, R., Nalbantoglu, J., Shaw, I., Kascsak, R.J.,

Bolton, D.C.,and Bendheim, P.E. Cellular isoform of the scrapie prion participates in lymphocyte activation. Cell 61, 185, 1990.

263. Hoops, T.C., and Rindler, M.J. Isolation of a cDNA encoding

glycoprotein-2 (GP-2), the major zymogen granule membrane protein: homology to uromodulin Tamm-Horsfall glycoprotein. J. Biol. Chem. 266, 4257, 1991.

264. Davis, S., Aldrich, T.H., Valenzuela, D.M., Wong, V., Furth, M.E.,

Squinto, S.P., and Yancopoulos, G.D. The receptor for ciliary neurotrophic factor. Science 253, 59, 1991

265. Seaman, W.E., Niemi, E.C., Stark, M.R., Goldfien, R.D., Pollock, A.S.,

and Imboden, J.B. Molecular cloning of GP42, a cell surface molecule that is selectively induced on rat natural-killer cells by interleukin-2:

Field and Menon 55

glycolipid membrane anchoring and capacity for transmembrane signalling. J. Exp. Med 173, 251, 1991.

266. Dhar, A., and Shukla, S.D. Release of a membrane surface

glycoprotein from human platelets by phosphatidylinositol specific phospholipase(s) C. Biochem. Biophys. Acta 1091, 15, 1991.

267. Corfield, A. P., Duggleby, S., Wagner, S.A., and Clamp, J.R. Human

colon membrane mucins have a glycosyl-phosphatidyl inositol membrane anchor. Biochem. Soc. Trans (London) 19, S220, 1991.

268. Wenger, R. H., Ayane, M., Bose, R., Kohler, G., and Nielsen, P.J. The

genes for a mouse hematopoietic differentiation marker called the heat stable antigen. Eur. J. Immunol. 21, 1039, 1991.

269. Early, A.E., Williams, T.G., Meyer, H.E., Por, S.R., Smith, E., et al.

Structural characterisation of Dictyostelium discoideum prespore-specific gene D19 and its product, cell surface glycoprotein PsA. Mol. Cell. Biol. 8, 3458, 1988.

270. Vai, M., Gatti, E., Lacana, E., Poplo, L., and Alberghina, L. Isolation and

deduced amino acid sequence of the gene encoding GP115, a yeast glycophospholipid-anchored protein containing a serine-rich region. J. Biol. Chem. 266, 12242, 1991.

271. Rand, K.N., Moore, T., Sriskantha, S., Spring, K., Tellam, R.,

Willadesen, P., and Cobon, G.S.Cloning and expression of a protective antigen from the cattle tick Boophilus microplus. Proc. Natl. Acad. Sci. USA 86, 9657, 1989.

272. Deddish, P.A., Skidgel, R.A., Kricho, V.B., Li, X-Y., Becker, R.P.,and

Erdos, E.G. Carboxypeptidase M in Madin-Darby canine cells. J. Biol. Chem. 265, 15083, 1990.

273. Sato, S., Zhu, X.L., and Sly, W.S. Carbonic anhydrase isozymes IV and

II in urinary membranes from carbonoc anhydrase II-deficient patients. Proc. Natl. Acad. Sci. USA 87, 6073, 1990.

274. Zhu, X.L., and Sly, W.S. Carbonic anhydrase IV from human lung. J.

Biol. Chem. 265, 8795, 1990. 275. Fischer, G.F., Majdic, O., Gadd, S., and Knapp, W. Signal transduction

in lymphocytic and myeloid cells via CD24, a new member of phosphoinositol-anchored membrane molecules. J. Immunol. 144, 638, 1990.

Field and Menon 56

276. Alberti, S., Miott, S., Fornaro, M., Mantovani, L., Walter, S., Canevari,

S., Menard, S., and Colnaghi, M.I. The CA-MOv18 molecule, a cell-surface marker of human ovarian carcinomas, is anchored to the cell membrane by phosphatidylinositol. Biochem. Biophys. Res. Comm. 171, 1051, 1990.

277. Cabrero, J., G., and Reiser, H. The sgp-60 molecule is linked to the

plasma membrane via a glycosylphosphatidylinositol anchor. Eur. J. Immunol. 21 1899, 1991.

278. Kubota, H., Okazaki, H., Onuma, M., Kano, S., Hattori, M., and Minato,

N.Identification and gene cloning of a new phosphatidylinositol-linked antigen expressed on mature lymphocytes. J. Immunol. 145, 3924, 1990.

279. Carroll, P.T., and Smith, L.K. Effects of phospholipase C from Bacillus

cereus on the release of membrane-bound choline-O-acetyltransferase from rat hippocampal tissue. J. Neurochem. 54, 1047, 1990.

280. Yoshihara, Y., Olea, S., Watanabe, Y., Mori, K. Developmentally and

spatially regulated expression of HNK-1 carbohydrate antigen on a novel phosphatidylinositol-anchored protein in rat brain. J. Cell Biol. 115, 731, 1991.

281. Ferguson, M.A.J. Lipid anchors on membrane proteins. Curr Op.

Struct. Biol. 1, 522 (1991). 282. Luhrs, C.A., and Slomiany, B.L. A human membrane-associated folate

binding protein is anchored by a glycosylphosphatidylinositol tail. J. Biol. Chem. 264, 21446, 1989.

283. Weinhart, U., Thomas, J.R., Pak, Y. Thompson, G.A., and Ferguson,

M.A.J. Structural characterization of a novel glycosylphosphatidylinositol from the protozoan Tetrahymena mimbres. Biochem. J. 279, 605, 1991.

284. Rosen, G., Pahlsson, P., Londner, M.V., Westerman, M.E., and Nilsson,

B. Structural analysis of glycophosphatidylinositol antigens of Leishmania major. J. Biol. Chem. 264, 10457, 1989.

Field and Menon 57

Figure Legends Figure 1; Isolation of VSG GPI-glycan fragment for structural analysis. The figure is redrawn from Ferguson et al. (1988)50, and shows the route to production of a neutral glycan from the GPI-anchor suitable for structural analysis. See text for a detailed discussion. Abbreviations used are; CO; carboxyl-terminus of the VSG polypeptide, EthN; ethanolamine, P; phosphate, P'; cyclic phosphodiester, GlcN; glucosamine, PI; phosphatidylethanolamine, DAG; diacylglycerol, AA; amino acid remaining following exhaustive pronase digestion, HONO; nitrous acid, AHM; anhydromannitol. Structures in bold are 3H-labelled. Figure 2; Schematic diagram of a GPI-anchor or lipid to show the positions of action of various reagents. The shaded region represents the neutral glycan core, which can be studied by liquid chromatography or exoglycosidase digestion. The figure is redrawn from Field and Menon (1992)88. See text for detailed discussion of individual reactions. R1 in GPI-anchored proteins is the C-terminal amino acid, and in precursor lipids is H. R4 is either H or palmitate. Figure 3; Attachment of a GPI-anchor to protein. The reaction probably occurs in the lumenal leaflet of the endoplasmic reticulum (ER) membrane bilayer, after translation of the protein on a membrane-bound ribosome, translocation of the polypeptide across the ER membrane, and cleavage of the the N-terminal signal sequence by ER signal peptidase. GPI attachment is signalled by a COOH-terminal peptide that is removed and replaced by a GPI, probably via a transpeptidation reaction mechanism. The GPI anchor precursor is shown to be inositol acylated (i.e., PI-PLC-resistant) as in mammalian cells. In some cells, the structure is processed (*) to a PI-PLC-sensitive form after attachment to protein, possibly after the protein leaves the ER (see section IV.3 for a complete discussion). Figure 4; Biosynthesis of GPI. The figure shows an outline of GPI biosynthesis indicating the immediate identified donors (boxed) of individual components of the structure, and sites of action (❄, ❆, ❇) of various inhibitors. Inositol acylated (PI-PLC resistant) structures are not shown, nor is the membrane topology of the assembly process (see sections IV.3 and IV.4 for details).

Table 1: PROTEINS WITH A GLYCOSYLPHOSPHATIDYLINOSITOL ANCHOR. The list does not necessarily include all proteins that are believed to utilise the GPI anchor. Note that the data for some molecules are more rigourous than others. We have included those proteins which are believed to be GPI-anchored solely on the basis of primary structure, as this criterion is becoming more reliable. The list includes those from the review by Low2. These proteins are not referenced for reasons of brevity, but references for all proteins additional to the 1989 compilation are provided. _____________________________________________________________________________________________________ Protein Source Reference _____________________________________________________________________________________________________ Protozoal antigens Ssp-4 T. cruzi 90kDa glycoprotein T. cruzi gp50-55 T. cruzi 238 gp85 family T. cruzi 239-242 Fl-160 (160kDal) T. cruzi 243 SAPA (shed acute phase antigen) T. cruzi 244 Sialidase (gp120-200kdal complex) T. cruzi 245 Variant surface glycoprotein T. brucei Procyclin/PARP T. brucei Procyclic Sialidase T. brucei 246 Transferrin binding protein [ESAG6] T. brucei 247

Surface antigens P. primaurelia p76 proteinase P. falciparum merizoites 195kDal antigen P. falciparum Transferrin receptor P. falciparum P30 T. gondii gp42/M-2 L. amazonensis 248 gp63 (PSP) Leishmania major, mexicana 249 PSA-2 L. major 250 GP49 Giardia lamblia 251 Cell-Cell Interactions LFA-3 human blood cells Leu8/TQ1 human endothelia 252 Heparan sulphate proteoglycan rat hepatocytes Neural cell adehesion molecule mammalian, chicken brain and muscle Fibronectin receptor human and avian tissues 253 TAG-1 mammalian neurons 254, 254b Opioid binding protein mammalian neural tissue 255 PH-20 guinea pig sperm F11 chicken brain

Fasciculin I D. melanogaster 256 Chaoptin D. melanogaster 257 AGα1 S. cerevisiae 258 AGA1 S. cerevisiae 259 Contact site A D. discoideum Miscellaneous Decay accelerating factor (DAF) human blood cells Urokinase Receptor mammalian tissues 260 NB-1 murine neutrophils 261 130kDal hepatoma glycoprotein rat hepatoma 34kDal growth factor human placental trophoblast scrapie prion (PrPSc, PrPC) mammalian brain, lymphocytes 262 GP-2 pig, rat zymogen granules Tamm-Horsfall glycoprotein mammalian tissues (kidney) 263 (Uromodulin) CD-16 (Fcγ Receptor III) human neutrophils Oligodendrocyte-myelin protein human brain CNTF Receptor mammalian brain 264 C8 binding protein human lymphocytes and monocytes

GP42 rat NK cells 265 Folate binding protein rat kidney, KB cells 26kDal glycoprotein rat sperm 150kDal glycoprotein [GPIb] human platelets 266 high molecular weight mucins human colon 267 82, 68kDal glycoproteins bovine adrenal chromaffin granule I-antigenic glycoprotein GP-3 calf thyroid Heat stable antigen murine hematopoietic lineage 268 Sgp-1, Sgp-2 squid neural tissue Surface antigens S. mansoni Antigen 117 D. discoideum PsA D. discoideum 269 125kDal glycoprotein [GAS1] S. cerevisiae 115kDal glycoprotein S. cerevisiae 270 Protective antigen B. microplus 271 Hydrolytic enzymes (non-protozoan) Alkaline phosphatase (APase) mammalian, S. mansonii, salamander

5'-Nucleotidase [CD73] mammalian Acetylcholinesterase T. marmorata, insect brain mammalian erythrocytes, Schistosomula Trehalase mammalian Alkaline phosphodiesterase I rat tissue Dipeptidase pig kidney, sheep lung Aminopeptidase P pig and human kidney Lipoprotein lipase 3T3-L1 adipocytes Carboxypeptidase M MDCK cells 272 Carbonic anhydrase II and IV human kidney 273,274 Mammalian antigens Thy-1 brain, lymphocytes Thy-3 murine lymphocytes RT-6 rat lymphocytes Qa murine lymphocytes Ly-6 (complex) murine lymphocytes MEM-43 human leukocytes and erythrocytes Carcinoembryonic antigen human tumour cells NCA human

Blast-1 human lymphocytes MRC OX-45 rat lymphocytes CD14 human monocytes Mo3 human monocytes CD48 human leukocytes CD24 human granulocytes 275 CA-MOv18 (38kDal antigen) human ovarian carcinoma 276 sgp-60 murine T and B cells 277 B7 (12kDal antigen) mammalian lymphocytes 278 Choline-O-Acetyltransferase rat hippocampus 279 PI-GP150 rat telencephalon 280 ______________________________________________________________________________________________________

Table 2: Structures of Protein GPI-anchors and their Precursors (ia) Proteins; Core Glycans and Modifications _______________________________________________________________________________________ Name Structure Reference Trypanosoma brucei VSG 117 Asp-E-P-Mα1-2Mα1-6Mα1-4GlcNα1-6PI☎ 50 ❘ +/-Galα1-2(+/-Galα1-2Galα1-6)Galα1-3 Trypanosoma brucei VSG 118 Asn-E-P-Mα1-2Mα1-6Mα1-4GlcNα1-6PI 46 Trypanosoma brucei VSG 221 Ser-E-P-Mα1-2Mα1-6Mα1-4GlcNα1-6PI 46 (+ 8 α-Gal, not fully defined) Trypanosoma cruzi 1G7 E-P-Mα1-2Mα1-6Mα1-4GlcNα1-6PI 281 ? ❘ Mα1 Leishmania major gp63 Asn-E-P-Mα1-2Mα1-6Mα1-4GlcNα1-6PI 53 (Acid protease) Saccharomyces cerevisiae gp125 E-P-Mα1-2Mα1-6Mα1-4GlcNα1-6PI 281 2 ❘ Mα1

Human erythrocyte Acetyl- Gly-E-P-Mα1-2Mα1-6Mα1-4GlcNα1-6PI 51,52 cholinesterasea 2 ❘ E-P

Rat Brain Thy-1 Cys-E-P-Mα1-2Mα1-6Mα1-4GlcNα1-6PI 64 2 2 4 ❘ ❘ ❘ Mα1 E-P GalNAcβ1 ________________________________________________________________________________________ The conserved core structure for the protein-linked GPI series is shown in bold. a. Linkage positions and the identity of the hexoses in the human acetylcholinesterase not determined, but assumed to be homologous. Abbreviations M = mannose, E = ethanolamine, P = phosphate. ☎ = Phosphatidylinositol (PI) = myoInositol1-P-CH2 ? ❘ ❘ CH-R1 R3 ❘ CH2-R2

Linkage position of the R3 substituent has not been determined.

(ib) Proteins; Lipid Substituents Protein R1 R2 R3 Reference ___________________________________________________________________________________________ Trypanosoma brucei VSG C14:0 C14:0 H 44 Trypanosoma brucei PARP C18:0 H C16:0 55 Leishmania major gp63a C18:0-23:0, C16:0, C14:0 H 53 C24:0, C25:0-26:0 C12:0, C18:0 Human erythrocyte AChEa C18:0, C18:1 C22:4,C22:5 C16:0 51,52 C22:6 Torpedo AChE C16:0 C16:0 ND 129 Human Folate Binding Proteina C20:0 C22:0 ND 252 ____________________________________________________________________________________________ a For these proteins, substituent at R1 is in ether linkage. Major species shown in bold. ND= not determined

(ii) GPI-Anchor Precursors Only mature precursors that have been well identified by chromatography and analytical methods are included in this table. A large number of other species have been identified, and are discussed in the text under biosynthesis. GPIa Protein Organism Lipid Substituents sn-1 sn-2 Inositol Reference _____________________________________________________________________________________________ P2 VSG T. brucei BSF C14:0 C14:0 - 59 89 P3 VSG? T. brucei BSF C14:0 C14:0 C16:0 59, 89 PP1 PARP T. brucei Procyclic C18:0 H C16:0 60 LP-1b gp63 L. major Not determined 76 Promastigote Lipid C Various Murine T cells Not determinedc 61,63 d Various Tetrahymena Not determined 283 ______________________________________________________________________________________________ a. Data are consistent with all the precursors containing the core ethanolaminylphosphoglycan. All linkages have not been determined for P2, P3 and PP1, but are assumed to be the same based on co-chromatography on HPLC. For LP-1, HPLC co-chromatography data available only. b. Lipid composition not determined, but substituent at sn-1 is base stable, and proposed to be alkyl, and homologous with the gp63 GPI-anchor. LP-1 is PI-PLC sensitive, and therefore does not contain a fatty acid inositol substituent. c. Lipid composition not determined, but resistance to PI-PLC indicates the probable presence of a fatty acid substituent on the inositol. d. The glycan sequence of this lipid is different from all other putative GPI-precursors. The Manα1-6Man linkage is replaced by Manα1-3Man. This lipid has not been definitively shown to be a GPI-precursor.

Table 3; Structures of the Lipophosphoglycan and Glycosylinositolphospholipid Series GPI Molecules The conserved core structure for the LPG/GIPL GPI series is shown in bold. Note that this sequence is also common to the protein GPI-anchor series. (i) Glycans (a) Lipophosphoglycan (LPG) Structure; Mα1-2Mα1-[P-6Gβ1-4Mα1-]nP-6Gα1-6Gα1-3Gfβ1-3Mα1-3Mα1-4GlcNα1-6PI* 4 3 6 ❘ ❘ ❘

+/-Gβ1 R +/-Glcα1-P Species R = Reference _____________________________________________________________________ L. donovani H 71 L. mexicana H or Glcβ1- 73 L. major H, [Gβ1-3]0-3Gβ1-, or Arap1-2[Gβ1-3]0-2Gβ1- 73 ______________________________________________________________________

Lipopetidophosphoglycan (LPPG) AEP 6 Structure; Gfβ1-3Mα1-2Mα1-2Mα1-6Mα1-4GlcN-Inositol-PO4-Ceramide 3 Gf1 Major surface glycoconjugate from epimastigotes of Trypanosoma cruzi Data from 77,78. Note that for this molecule, homology to the protein-anchor series is more extensive and includes the first three mannose residues attached to the GlcN. n (average number of repeats) = 10-30. * PI in all cases is a lyso-alkylPI AEP = 2-aminoethylphosphonate.

(b) Glycoinositolphospholipids (GIPLs) L. major promastigote. Core structure; R-Gfα1-3Mα1-3Mα1-4GlcNα1-6PI Lipid Species R = Abundance (Strain)/Mol% V121 L119 ______________________________________________________________ GIPL-1 H 9 2 GIPL-2a Gα1-3 33 26 GIPL-3 Gα1-6Gα1-3 26 8 GIPL-A Gα1-3Gα1-3 4 ND GIPL-4, 5, 6 NDb ND + ______________________________________________________________ L. donovani. Core structure; R'-6[R''-3]Mα1-4GlcNα1-6PI Lipid Species R' R'' __________________________________________________________ Promastigote stage IsoM2 H Mα1- IsoM3 Mα1- Mα1- IsoM4 Mα1-2Mα1- Mα1- Amastigote stage M2 Mα1- H M3 Mα1-2Mα1- H __________________________________________________________

a Lyso forms of GIPLs 2 and 3 also present. b GIPLs 4-6 detected in the L119 strain, and possibly LPG precursors. ND Not determined. Data from 67, 68, 69, 73.

(ii) Lipid Moieties Structure/Spp sn-1(Alkyl) sn-2(Acyl) Ref ___________________________________________________________________________ L.major GPI A,B,C/ C24:0, C20:0, C18:0, C16:0, 283 GIPL1-4 C18;0, C16:0 C14:0 67 GIPL5,6 C24:0, C18:0, Not present 67 C26:0 LPG C24:0, C26:0 Not present 73 L. donovani M2,3, IsoM2,3,4 C18:0 C18:0, C16:0 69 C14:0 LPG C24:0, C26:0 Not present 57 _______________________________________________________________________________________ Major species are shown in bold.